qat

Step-by-Step

This document is used to list steps of reproducing TensorFlow keras Intel® Neural Compressor QAT conversion.

Prerequisite

1. Installation

# Install Intel® Neural Compressor
pip install neural-compressor

2. Install Intel Tensorflow and TensorFlow Model Optimization

pip install intel-tensorflow==2.4.0
pip install tensorflow_model_optimization==0.5.0

Note: To generate correct qat model with tensorflow_model_optimization 0.5.0, pls use TensorFlow 2.4 or above.

3. Prepare Pretrained model

Run the train.py script to get pretrained fp32 model.

4. Prepare QAT model

Run the qat.py script to get QAT model which in fact is a fp32 model with quant/dequant pair inserted.

Run Command

python convert.py     # to convert QAT model to quantized model.

python benchmark.py   # to run accuracy benchmark.