ONNX helper scripts

This repo contains small helper scripts for working with ONNX files

Installation

Setup the environment for utility scripts, by creating a virtual environment and installing dependencies Note: Tested using python3.8. Run the following commands:

git clone https:/github.com/rahul-tuli/onnx-helpers
cd onnx-helpers
python3.8 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Scripts

committed- Sample Out Generator A script to generate sample outputs from sample inputs and a given ONNX file, supports both DeepSparse and ONNX Runtime Engine. Inspired from an original version by Ben Fineran. Find more about the usage by python3 sample_out_generator.py --help

Examples:

❯ python sample_out_generator.py --help
usage: Generate sample outputs and from inputs [-h] --model-path MODEL_PATH --sample-inputs SAMPLE_INPUTS [--save-dir SAVE_DIR]
                                               [--engine {ort,deepsparse}]

optional arguments:
  -h, --help            show this help message and exit
  --model-path MODEL_PATH
                        ONNX model filepath
  --sample-inputs SAMPLE_INPUTS
                        Directory containing sample inputs
  --save-dir SAVE_DIR   Directory to save the sample outs, copy inputs and model. Defaults to model directory
  --engine {ort,deepsparse}
                        Engine to use for generating outs

python sample_out_generator --model-path ONNX_FILE \
--sample-inputs INPUTS_DIRECTORY \
--save-dir SAVE_DIR \
--engine deepsparse