AIDGE Post Training Quantization tutorial

Description

This MR proposes a tutorial for performing Post Training Quantization of a model through the aidge_quantization module of the python API. After setting up the environment for inferences on the MNIST dataset, PTQ is performed and accuracies are compared before and after quantization. In the case of 8-bit PTQ, we do not observe any accuracy degradation on the considered set of samples.

Added files/folders

• ptq_tuto.ipynb, the jupyter notebook containing the tutorial
• ptq_diagram.png, the schematic describing the PTQ pipeline
• assets/, the folder containing the numpy samples and the onnx model