├── .gitignore ├── .gitmodules ├── LICENCE ├── README.md ├── csrc ├── CMakeLists.txt ├── cutlass_kernel_file_1.generated.cu ├── cutlass_kernel_file_2.generated.cu └── w2a16.cu ├── datautils.py ├── decoupleQ ├── __init__.py ├── linear_w2a16.py ├── moq_quant.py └── quant.py ├── imgs ├── img.png └── private_exp.png ├── llama.py ├── requirements.txt ├── resnet.py ├── run_inference_llama.sh ├── run_llama.sh └── run_resnet.sh /.gitignore: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/.gitignore -------------------------------------------------------------------------------- /.gitmodules: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/.gitmodules -------------------------------------------------------------------------------- /LICENCE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/LICENCE -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/README.md -------------------------------------------------------------------------------- /csrc/CMakeLists.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/csrc/CMakeLists.txt -------------------------------------------------------------------------------- /csrc/cutlass_kernel_file_1.generated.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/csrc/cutlass_kernel_file_1.generated.cu -------------------------------------------------------------------------------- /csrc/cutlass_kernel_file_2.generated.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/csrc/cutlass_kernel_file_2.generated.cu -------------------------------------------------------------------------------- /csrc/w2a16.cu: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/csrc/w2a16.cu -------------------------------------------------------------------------------- /datautils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/datautils.py -------------------------------------------------------------------------------- /decoupleQ/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /decoupleQ/linear_w2a16.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/decoupleQ/linear_w2a16.py -------------------------------------------------------------------------------- /decoupleQ/moq_quant.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/decoupleQ/moq_quant.py -------------------------------------------------------------------------------- /decoupleQ/quant.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/decoupleQ/quant.py -------------------------------------------------------------------------------- /imgs/img.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/imgs/img.png -------------------------------------------------------------------------------- /imgs/private_exp.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/imgs/private_exp.png -------------------------------------------------------------------------------- /llama.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/llama.py -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/requirements.txt -------------------------------------------------------------------------------- /resnet.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/resnet.py -------------------------------------------------------------------------------- /run_inference_llama.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/run_inference_llama.sh -------------------------------------------------------------------------------- /run_llama.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/run_llama.sh -------------------------------------------------------------------------------- /run_resnet.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ByteDance-Seed/decoupleQ/HEAD/run_resnet.sh --------------------------------------------------------------------------------