Commit History

作者 SHA1 備註 提交日期
  AlpinDale f561a54a43 core: fix async postprocessor in case of preemption (#1000) 1 月之前
  AlpinDale df5bf69938 core: slightly improve chunked prefill performance (#981) 1 月之前
  AlpinDale 5be6225f38 core: support multi-step scheduling w/ async post-processor (#955) 1 月之前
  AlpinDale 0e2bfccda0 core: add virtual engine for async outproc (#939) 1 月之前
  AlpinDale f7f3fed265 feat: add async postprocessor (#925) 1 月之前
  AlpinDale abfd4465ca feat: add support for chunked prefill + prefix caching (#871) 1 月之前
  AlpinDale 3d83e64f8e feat: add metrics for prefix cache hit rate (#829) 2 月之前
  AlpinDale 2f61644f6e SPMD optimizations (#824) 2 月之前
  AlpinDale 0a369f9171 feat: support chunked prefill with LoRA (#823) 2 月之前
  AlpinDale 577586309d chore: multi-step args and sequence modifications (#713) 4 月之前
  AlpinDale bf88c8567e feat: mamba model support (#674) 4 月之前
  AlpinDale 7df7b8ca53 optimization: reduce end-to-end overhead from python obj allocation (#666) 4 月之前
  AlpinDale a0e446a17d feat: initial encoder-decoder support with BART model (#633) 5 月之前
  AlpinDale f1d0b77c92 [0.6.0] Release Candidate (#481) 5 月之前
  AlpinDale 9d81716bfd [v0.5.3] Release Candidate (#388) 8 月之前
  AlpinDale 78d66f16d1 Chunked Prefill Part 1 (#384) 10 月之前
  AlpinDale f8dfac6372 chore: attention refactor and upstream sync apr01 (#365) 10 月之前
  AlpinDale e42a78381a feat: switch from pylint to ruff (#322) 10 月之前
  AlpinDale c2d77b1822 chore: logging refactor (#302) 10 月之前
  AlpinDale ac82b67f75 feat: naive context shift and various QoL changes (#289) 11 月之前
  AlpinDale 4d04ade9ef feat: fine-grained seeds (#279) 11 月之前
  AlpinDale c0aac15421 feat: S-LoRA support (#222) 1 年之前
  AlpinDale 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 1 年之前
  AlpinDale b9b295d74e chore: backlogs 1 (#191) 1 年之前
  g4rg 2aab3da9bd chore: fix Python 3.8+ compatibility (#170) 1 年之前
  AlpinDale 9ec4e08ade fix: cpu sync delay fix (#127) 1 年之前
  AlpinDale 13901af940 fix: scheduler hang with long prompts (#126) 1 年之前
  50h100a fa0ae5a2c9 feat: new mirostatv2 implementation (#96) 1 年之前
  AlpinDale efc6f7fbec chore: reformats (#90) 1 年之前
  AlpinDale 3d72f05c7b feat: flattened 1D tensor -> 2D tensor (#85) 1 年之前