1
0

Коммит түүх

Эзэн SHA1 Мессеж Огноо
  AlpinDale 60ca1e1e5e feat: add ngram prompt lookup decoding for speculative decoding (#438) 9 сар өмнө
  AlpinDale d8c4193704 feat: Speculative Decoding using a draft model (#432) 9 сар өмнө
  AlpinDale 76f36af704 feat: LM Format Enforcer support (#428) 9 сар өмнө
  AlpinDale bd0ddf1cfe feat: EETQ quantization (#408) 9 сар өмнө
  AlpinDale 4d33ce60da feat: Triton flash attention backend for ROCm (#407) 9 сар өмнө
  AlpinDale 071269e406 feat: FP8 E4M3 KV Cache (#405) 9 сар өмнө
  AlpinDale 6f00203041 refactor scheduler for chunked prefill, remove reorder policy for now 9 сар өмнө
  AlpinDale 9aaeb5d349 add speculative config and arg for later 9 сар өмнө
  AlpinDale a304f76d89 feat: Intel CPU support (#403) 9 сар өмнө
  AlpinDale fa083286e3 Speculative Decoding Part 4: Lookahead scheduling (#402) 9 сар өмнө
  AlpinDale f845a661dd Chunked Prefill Part 2: data update 10 сар өмнө
  AlpinDale eff5eb16c5 ruff 10 сар өмнө
  AlpinDale 753f6dc51b add v2 block manager 10 сар өмнө
  AlpinDale 7b9c08afae vision model support 10 сар өмнө
  AlpinDale b738554558 add reorder scheduler policy 10 сар өмнө
  AlpinDale 1ba9ff78cd add scheduler delay factor 10 сар өмнө
  AlpinDale 78d66f16d1 Chunked Prefill Part 1 (#384) 10 сар өмнө
  AlpinDale feb5840f2a feat: async tokenization (#374) 10 сар өмнө
  AlpinDale f8dfac6372 chore: attention refactor and upstream sync apr01 (#365) 10 сар өмнө
  AlpinDale c41462cfcd feat: exllamav2 quantization (#305) 10 сар өмнө
  AlpinDale 9810daa699 feat: INT8 KV Cache (#298) 11 сар өмнө
  AlpinDale e0c35bb353 feat: bitsandbytes and `--load-in{4,8}bit` support (#294) 11 сар өмнө
  AlpinDale 705821a7fe feat: AQLM quantization support (#293) 11 сар өмнө
  AlpinDale ac82b67f75 feat: naive context shift and various QoL changes (#289) 11 сар өмнө
  AlpinDale 72229a94da feat: better marlin kernels (#285) 11 сар өмнө
  AlpinDale 657aec0cbd refactor: OpenAI endpoint (#261) 11 сар өмнө
  AlpinDale 4d04ade9ef feat: fine-grained seeds (#279) 11 сар өмнө
  AlpinDale ea0f57b233 feat: allow further support for non-cuda devices (#247) 1 жил өмнө
  AlpinDale c3a221eb02 feat: GGUF, QuIP#, and Marlin support (#228) 1 жил өмнө
  AlpinDale 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 жил өмнө