The ModelAccuracyTool-Armnn is a program for measuring the Top 5 accuracy results of a model against an image dataset.
Prerequisites:
ArmnnConverter can be used to convert a model to this format.Build option: To build ModelAccuracyTool, pass the following options to Cmake:
| Cmd: | ||
|---|---|---|
| -h | --help | Display help messages |
| -m | --model-path | Path to armnn format model file |
| -f | --model-format | The model format. Supported values: caffe, tensorflow, tflite |
| -i | --input-name | Identifier of the input tensors in the network separated by comma |
| -o | --output-name | Identifier of the output tensors in the network separated by comma |
| -d | --data-dir | Path to directory containing the ImageNet test data |
| -p | --model-output-labels | Path to model output labels file. |
| -v | --validation-labels-path | Path to ImageNet Validation Label file |
| -l | --data-layout ] | Data layout. Supported value: NHWC, NCHW. Default: NHWC |
| -c | --compute | Which device to run layers on by default. Possible choices: CpuRef, CpuAcc, GpuAcc. Default: CpuAcc, CpuRef |
| -r | --validation-range | The range of the images to be evaluated. Specified in the form :. The index starts at 1 and the range is inclusive. By default the evaluation will be performed on all images. |
| -b | --blacklist-path | Path to a blacklist file where each line denotes the index of an image to be excluded from evaluation. |
Example usage:
./ModelAccuracyTool -m /path/to/model/model.armnn -f tflite -i input -o output -d /path/to/test/directory/ -p /path/to/model-output-labels -v /path/to/file/val.txt -c CpuRef -r 1:100