Block Floating Point (BFP) and DSConv with GPU support

This repo implements BFP and DSConv in cuda kernels to be used with PyTorch

Requirements

Download CUB and put it in /home/your_username/libs/ (or the file indicated at NUQ/BlackBox/Quantization/src/setup.py:22)
cd /path/to/NUQ/BlackBox/Quantization/src/ then python build_ext --inplace.

Name		Name	Last commit message	Last commit date
Latest commit History 82 Commits
DSConv/nn		DSConv/nn
cifar_models		cifar_models
models		models
src		src
.gitignore		.gitignore
README.md		README.md
__init__.py		__init__.py
base_model.py		base_model.py
test_case_activation.py		test_case_activation.py
test_models.py		test_models.py
test_state_dict.py		test_state_dict.py
train_CIFAR10.py		train_CIFAR10.py