| /// Copyright (c) 2020 ARM Limited. |
| /// |
| /// SPDX-License-Identifier: MIT |
| /// |
| /// Permission is hereby granted, free of charge, to any person obtaining a copy |
| /// of this software and associated documentation files (the "Software"), to deal |
| /// in the Software without restriction, including without limitation the rights |
| /// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell |
| /// copies of the Software, and to permit persons to whom the Software is |
| /// furnished to do so, subject to the following conditions: |
| /// |
| /// The above copyright notice and this permission notice shall be included in all |
| /// copies or substantial portions of the Software. |
| /// |
| /// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| /// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| /// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| /// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| /// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| /// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| /// SOFTWARE. |
| /// |
| |
| namespace armnn |
| { |
| /** |
| @page converter_and_quantizer The ArmNN Converter and Quantizer |
| @tableofcontents |
| |
| @section S10_converter The ArmNN Converter |
| |
| The `ArmnnConverter` is a program for converting neural networks from other formats to Arm NN format. |
| Currently the program supports models in Caffe, Onnx, Tensorflow Protocol Buffers and Tensorflow Lite FlatBuffers formats. Run the program with no arguments to see command-line help. |
| |
| For more information about the layers that are supported, see <a href="parsers.xhtml">parsers</a>. |
| <br/><br/><br/><br/> |
| |
| @section S11_quantizer The ArmNN Quantizer |
| |
| The `ArmnnQuantizer` is a program for loading a 32-bit float network into ArmNN and converting it into a quantized asymmetric 8-bit or quantized symmetric 16-bit network. |
| It supports static quantization by default, dynamic quantization is enabled if CSV file of raw input tensors is provided. Run the program with no arguments to see command-line help. |
| |
| |
| |Cmd:||| |
| | ---|---|---| |
| | -h | --help | Display help messages | |
| | -f | --infile | Input file containing float 32 ArmNN Input Graph | |
| | -s | --scheme | Quantization scheme, "QAsymm8" or "QSymm16". Default value: QAsymm8 | |
| | -c | --csvfile | CSV file containing paths for raw input tensors for dynamic quantization. If unset, static quantization is used | |
| | -p | --preserve-data-type | Preserve the input and output data types. If unset, input and output data types are not preserved | |
| | -d | --outdir | Directory that output file will be written to | |
| | -o | --outfile | ArmNN output file name | |
| |
| <br/> |
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.sh |
| ./ArmnnQuantizer -f /path/to/armnn/input/graph/ -s "QSymm16" -c /path/to/csv/file -p 1 -d /path/to/output -o outputFileName |
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
| |
| **/ |
| } |