├── .DS_Store ├── .gitignore ├── LICENSE ├── README.md ├── bitdelta ├── binary_gemm_kernel.py ├── data.py ├── diff.py ├── eval_ppl.py ├── misc.py ├── train.py └── utils.py ├── demo ├── README.md ├── demo_backend.py ├── demo_gradio.py └── supported_models.json ├── docs ├── .gitignore ├── index.html └── static │ ├── css │ ├── bulma-carousel.min.css │ ├── bulma-slider.min.css │ ├── bulma.css.map.txt │ ├── bulma.min.css │ ├── fontawesome.all.min.css │ └── index.css │ ├── images │ ├── BitDelta.png │ ├── kernel_batch_size.png │ ├── kernel_hidden_size.png │ └── nbit.png │ └── js │ ├── bulma-carousel.js │ ├── bulma-carousel.min.js │ ├── bulma-slider.js │ ├── bulma-slider.min.js │ ├── fontawesome.all.min.js │ └── index.js ├── figures ├── BitDelta.png └── BitDeltaDemo.webm ├── notebooks ├── binary_gemm_kernel_triton.ipynb ├── compression_lora.ipynb ├── compression_ternary.ipynb ├── mixtral_weight.ipynb └── multilingual_eval.ipynb ├── pyproject.toml └── scripts ├── multigpu_train_example.bash ├── ppl_eval_example.bash └── train_example.bash /.DS_Store: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/.DS_Store -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/.gitignore -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/LICENSE -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/README.md -------------------------------------------------------------------------------- /bitdelta/binary_gemm_kernel.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/binary_gemm_kernel.py -------------------------------------------------------------------------------- /bitdelta/data.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/data.py -------------------------------------------------------------------------------- /bitdelta/diff.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/diff.py -------------------------------------------------------------------------------- /bitdelta/eval_ppl.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/eval_ppl.py -------------------------------------------------------------------------------- /bitdelta/misc.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/misc.py -------------------------------------------------------------------------------- /bitdelta/train.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/train.py -------------------------------------------------------------------------------- /bitdelta/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/bitdelta/utils.py -------------------------------------------------------------------------------- /demo/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/demo/README.md -------------------------------------------------------------------------------- /demo/demo_backend.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/demo/demo_backend.py -------------------------------------------------------------------------------- /demo/demo_gradio.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/demo/demo_gradio.py -------------------------------------------------------------------------------- /demo/supported_models.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/demo/supported_models.json -------------------------------------------------------------------------------- /docs/.gitignore: -------------------------------------------------------------------------------- 1 | .DS_store 2 | .idea 3 | -------------------------------------------------------------------------------- /docs/index.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/index.html -------------------------------------------------------------------------------- /docs/static/css/bulma-carousel.min.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/bulma-carousel.min.css -------------------------------------------------------------------------------- /docs/static/css/bulma-slider.min.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/bulma-slider.min.css -------------------------------------------------------------------------------- /docs/static/css/bulma.css.map.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/bulma.css.map.txt -------------------------------------------------------------------------------- /docs/static/css/bulma.min.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/bulma.min.css -------------------------------------------------------------------------------- /docs/static/css/fontawesome.all.min.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/fontawesome.all.min.css -------------------------------------------------------------------------------- /docs/static/css/index.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/css/index.css -------------------------------------------------------------------------------- /docs/static/images/BitDelta.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/images/BitDelta.png -------------------------------------------------------------------------------- /docs/static/images/kernel_batch_size.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/images/kernel_batch_size.png -------------------------------------------------------------------------------- /docs/static/images/kernel_hidden_size.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/images/kernel_hidden_size.png -------------------------------------------------------------------------------- /docs/static/images/nbit.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/images/nbit.png -------------------------------------------------------------------------------- /docs/static/js/bulma-carousel.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/bulma-carousel.js -------------------------------------------------------------------------------- /docs/static/js/bulma-carousel.min.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/bulma-carousel.min.js -------------------------------------------------------------------------------- /docs/static/js/bulma-slider.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/bulma-slider.js -------------------------------------------------------------------------------- /docs/static/js/bulma-slider.min.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/bulma-slider.min.js -------------------------------------------------------------------------------- /docs/static/js/fontawesome.all.min.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/fontawesome.all.min.js -------------------------------------------------------------------------------- /docs/static/js/index.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/docs/static/js/index.js -------------------------------------------------------------------------------- /figures/BitDelta.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/figures/BitDelta.png -------------------------------------------------------------------------------- /figures/BitDeltaDemo.webm: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/figures/BitDeltaDemo.webm -------------------------------------------------------------------------------- /notebooks/binary_gemm_kernel_triton.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/notebooks/binary_gemm_kernel_triton.ipynb -------------------------------------------------------------------------------- /notebooks/compression_lora.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/notebooks/compression_lora.ipynb -------------------------------------------------------------------------------- /notebooks/compression_ternary.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/notebooks/compression_ternary.ipynb -------------------------------------------------------------------------------- /notebooks/mixtral_weight.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/notebooks/mixtral_weight.ipynb -------------------------------------------------------------------------------- /notebooks/multilingual_eval.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/notebooks/multilingual_eval.ipynb -------------------------------------------------------------------------------- /pyproject.toml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/pyproject.toml -------------------------------------------------------------------------------- /scripts/multigpu_train_example.bash: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/scripts/multigpu_train_example.bash -------------------------------------------------------------------------------- /scripts/ppl_eval_example.bash: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/scripts/ppl_eval_example.bash -------------------------------------------------------------------------------- /scripts/train_example.bash: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/FasterDecoding/BitDelta/HEAD/scripts/train_example.bash --------------------------------------------------------------------------------