Skip to content

GitLab

Switch branch/tag
  1. 10 Feb, 2026 1 commit
    • zhuwenwen's avatar
      support fa kvcache fp8, add VLLM_USE_QUERY_QUANT to not use q quant(todo) · 0dfb30d5
      zhuwenwen authored 
      [opt] 优化epsp代码, 零消耗添加epsp
update VLLM_USE_FUSED_RMS_ROPE=0 (default). for qwen3, VLLM_USE_FUSED_RMS_ROPE=1 (default)
feat(moe/marlin): Marlin W16A16 MoE 自动探测并预打包（去掉手动开关）
perf(qwen3): 融合 q/k RMSNorm + RoPE
fused_moe_fp8接入lmslim
      0dfb30d5
  3. 09 Feb, 2026 1 commit
  5. 31 Dec, 2025 2 commits
  7. 30 Dec, 2025 2 commits
  9. 29 Dec, 2025 4 commits
  11. 22 Dec, 2025 2 commits
  13. 18 Dec, 2025 2 commits
  15. 09 Dec, 2025 1 commit
  17. 09 Oct, 2025 1 commit
  19. 24 Sep, 2025 1 commit
  21. 16 Sep, 2025 1 commit
  23. 04 Sep, 2025 1 commit
  25. 27 Aug, 2025 1 commit
  27. 25 Aug, 2025 2 commits
  29. 18 Aug, 2025 1 commit
  31. 16 Aug, 2025 1 commit
  33. 15 Aug, 2025 1 commit
  35. 11 Aug, 2025 1 commit