runtime/test/generated/spec_V1_2/vts_V1_2/channel_shuffle.model.cpp

// Generated from channel_shuffle.mod.py
// DO NOT EDIT
// clang-format off
#include "GeneratedTests.h"

namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_relaxed_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT32,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 12, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
        .relaxComputationFloat32toFloat16 = true,
    };
}

bool is_ignored_relaxed_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_quant8_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_QUANT8_ASYMM,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.25f,
            .zeroPoint = 128,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 1},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 9, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 128, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_quant8_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 252, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim4_axis3_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 2, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 253, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {2, 3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim3_axis2_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12, 3},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 254, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0, 0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {3, 12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim2_axis1_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_neg() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_neg(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_neg_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {0};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_neg_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_neg_all_inputs_as_internal() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_neg_all_inputs_as_internal(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle
namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle {

Model createTestModel_float16_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape() {
    const std::vector<Operand> operands = {
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::TEMPORARY_VARIABLE,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 0, .length = 4},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 4, .length = 4},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {0},
            .numberOfConsumers = 0,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_OUTPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {12},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::MODEL_INPUT,
            .location = {.poolIndex = 0, .offset = 0, .length = 0},
        },
        {
            .type = OperandType::TENSOR_FLOAT16,
            .dimensions = {1},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 8, .length = 2},
        },
        {
            .type = OperandType::INT32,
            .dimensions = {},
            .numberOfConsumers = 1,
            .scale = 0.0f,
            .zeroPoint = 0,
            .lifetime = OperandLifeTime::CONSTANT_COPY,
            .location = {.poolIndex = 0, .offset = 10, .length = 4},
        }
    };

    const std::vector<Operation> operations = {
        {
            .type = OperationType::ADD,
            .inputs = {4, 5, 6},
            .outputs = {0},
        },
        {
            .type = OperationType::CHANNEL_SHUFFLE,
            .inputs = {0, 1, 2},
            .outputs = {3},
        }
    };

    const std::vector<uint32_t> inputIndexes = {4};
    const std::vector<uint32_t> outputIndexes = {3};
    std::vector<uint8_t> operandValues = {
      3, 0, 0, 0, 255, 255, 255, 255, 0, 0, 0, 0, 0, 0
    };
    const std::vector<hidl_memory> pools = {};

    return {
        .operands = operands,
        .operations = operations,
        .inputIndexes = inputIndexes,
        .outputIndexes = outputIndexes,
        .operandValues = operandValues,
        .pools = pools,
    };
}

bool is_ignored_float16_dim1_axis0_neg_all_inputs_as_internal_dynamic_output_shape(int i) {
  static std::set<int> ignore = {};
  return ignore.find(i) != ignore.end();
}

} // namespace android::hardware::neuralnetworks::V1_2::generated_tests::channel_shuffle