"vscode:/vscode.git/clone" did not exist on "aca5c514876303913c686866cd7571638e512064"
  1. 10 Feb, 2026 1 commit
    • zhuwenwen's avatar
      [feat] 适配block和channel fp8, 添加dp attention功能 · 2b7b1a31
      zhuwenwen authored
      [pd] 修改非堆成切分的判断, pd分离使用默认调度,set VLLM_USE_PD_SPLIT=1
      [perf](fused-moe): 预打包 Marlin W16A16 MoE 权重,降低 warmup 显存峰值
      [fix]修复丢弃MTP代码报错,pp+chunkprefill多并发input ids更新bug, 修复不开启融合图的断言错误, PP 场景 decode 阶段 token 被误丢弃导致卡住
      2b7b1a31
  2. 06 Jan, 2026 1 commit
  3. 29 Dec, 2025 1 commit
  4. 25 Dec, 2025 2 commits
  5. 23 Dec, 2025 1 commit
  6. 22 Dec, 2025 2 commits
  7. 21 Dec, 2025 1 commit
  8. 18 Dec, 2025 3 commits
  9. 09 Dec, 2025 1 commit
  10. 05 Dec, 2025 1 commit
    • laibao's avatar
      [bugfix] 优化 reject-sampling 的 InputBatch 元数据处理 · a0d556fe
      laibao authored
      - 在 InputBatch.refresh_metadata 中为展开后的采样元数据引入 repeat_count 记录重复次数
      - 完善元数据刷新逻辑以适配 reject-sampling 优化路径
      - 更新 GPUModelRunnerBase,在 batch 处理阶段正确消费新的采样元数据与重复计数
      a0d556fe
  11. 28 Nov, 2025 1 commit
  12. 27 Nov, 2025 1 commit
  13. 26 Nov, 2025 2 commits
  14. 21 Nov, 2025 1 commit
  15. 15 Nov, 2025 1 commit
  16. 13 Nov, 2025 1 commit
  17. 27 Oct, 2025 1 commit
  18. 17 Oct, 2025 2 commits
  19. 13 Oct, 2025 1 commit
  20. 11 Oct, 2025 2 commits
  21. 10 Oct, 2025 3 commits
  22. 09 Oct, 2025 1 commit
  23. 30 Sep, 2025 1 commit
  24. 26 Sep, 2025 1 commit
  25. 25 Sep, 2025 1 commit
  26. 24 Sep, 2025 2 commits
  27. 16 Sep, 2025 1 commit
  28. 06 Sep, 2025 2 commits
  29. 05 Sep, 2025 1 commit