Quantization Polish
Description
This merge request aims to enhance drastically the quantization module, by reworking the code, documenting it, and fixing a critical bug. The new code should allow to run the quantization tutorial on multibranch architectures. Here is a list of the provided changes:
-
Simplify the handling of residual branches -
Fix a critical bug in the quantization algorithm -
Clean the headers and includes -
Replace the 'cout' logs by 'fmt' ones -
Add the doxygen documentation in the header -
Add the sphinx docstrings in the pybind -
Fix the import of fmt in the CMake -
Update the Readme.md
Modified files : QuantPTQ.hpp
, QuantPTQ.cpp
, pybind_QuantPTQ.cpp
, CMakeLists.txt
and Readme.md
Edited by Benjamin Halimi