Komit Sejarah

Pembuat SHA1 Pesan Tanggal
  AlpinDale c18bf116da fix stop strings not being excluded from outputs 10 bulan lalu
  AlpinDale fe17712f29 fully working chunked prefill 10 bulan lalu
  AlpinDale 4d33ce60da feat: Triton flash attention backend for ROCm (#407) 10 bulan lalu
  AlpinDale 082d4e6972 feat: add chunked prefill scheduler (#406) 10 bulan lalu
  AlpinDale 7528e0ce3e make detokenization optional 10 bulan lalu
  AlpinDale 071269e406 feat: FP8 E4M3 KV Cache (#405) 10 bulan lalu
  AlpinDale 6f00203041 refactor scheduler for chunked prefill, remove reorder policy for now 10 bulan lalu
  AlpinDale 9aaeb5d349 add speculative config and arg for later 10 bulan lalu
  AlpinDale a304f76d89 feat: Intel CPU support (#403) 10 bulan lalu
  AlpinDale f845a661dd Chunked Prefill Part 2: data update 10 bulan lalu
  AlpinDale 753f6dc51b add v2 block manager 10 bulan lalu
  AlpinDale 3f5ce50c19 add stop_reason 10 bulan lalu
  AlpinDale 7b9c08afae vision model support 10 bulan lalu
  AlpinDale 0c4ead5e9f min_tokens 10 bulan lalu
  AlpinDale d1786645a3 fix formatting 10 bulan lalu
  AlpinDale eed70dff76 improve detokenization performance; improve logprobs 10 bulan lalu
  AlpinDale 2319b411ce refactor: neuron support 10 bulan lalu
  AlpinDale c9cb00c2a1 add warning for mismatch in vocab size 10 bulan lalu
  AlpinDale feb5840f2a feat: async tokenization (#374) 10 bulan lalu
  AlpinDale 0f6d56b07f feat: model executor refactor (#367) 10 bulan lalu
  AlpinDale b361096463 fix: tokenizer when using ray (#366) 10 bulan lalu
  AlpinDale f8dfac6372 chore: attention refactor and upstream sync apr01 (#365) 10 bulan lalu
  AlpinDale e42a78381a feat: switch from pylint to ruff (#322) 11 bulan lalu
  AlpinDale c2d77b1822 chore: logging refactor (#302) 11 bulan lalu
  AlpinDale 9810daa699 feat: INT8 KV Cache (#298) 11 bulan lalu
  AlpinDale 2d3d44b3e9 chore: add health check for ray workers (#290) 11 bulan lalu
  AlpinDale ac82b67f75 feat: naive context shift and various QoL changes (#289) 11 bulan lalu
  AlpinDale 657aec0cbd refactor: OpenAI endpoint (#261) 11 bulan lalu
  AlpinDale 4b80b42362 fix: memory leaks due to nccl cuda graphs (#275) 11 bulan lalu
  AlpinDale 2c08aa5af4 chore: remove eos token from output (#272) 11 bulan lalu