Aiming to accelerate neural network inference time on embedded systems via model compression, specifically pruning and quantization.
# Set up Python environment
virtualenv --python $(which python3) venv
source venv/bin/activate
pip install -r requirements.txt
# Download dataset
sh download_dataset.sh
# classify image with unmodified VGG-16
python run_classify.py
# prune model
python prune_model.py