Vitis Quantizer

THIS IS NOT AN OFFICIAL XILINX PACKAGE

A pip installable package to do the vitis model quantization for tensorflow 2.

This is a migration of the code here to a standalone pip package that won't conflict with the tensorflow-model-optimization package.

Working in the docker image is annoying and this code should be standalone.

Install

pip install vitis-quantizer

Build from source:

python3 setup.py bdist_wheel
pip install dist/vitis_quantizer-0.1.0-py3-none-any.whl

Usage

Usage is the same as Vitis AI models.

import tensorflow as tf
import vitis_quantizer

# Train/Get/Make a keras model somehow
model = tf.keras.models.load_model("/path/to/keras/model")
quantizer = vitis_quantizer.VitisQuantizer(model)
with vitis_quantizer.quantize_scope():
    quantized_model = quantizer.quantize_model(calib_dataset=dataset)
    quantized_model.save("/path/to/save/quantized/model")

After you have the quantized model saved, use vitis compile.sh script.