- 10 Feb, 2026 1 commit
-
-
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
-
- 09 Feb, 2026 1 commit
-
-
王敏 authored
-
- 31 Dec, 2025 4 commits
- 30 Dec, 2025 2 commits
- 29 Dec, 2025 6 commits
- 24 Dec, 2025 2 commits
- 22 Dec, 2025 2 commits
- 18 Dec, 2025 2 commits
- 09 Dec, 2025 1 commit
-
-
王敏 authored
-
- 27 Nov, 2025 1 commit
-
-
王敏 authored
-
- 26 Nov, 2025 1 commit
-
-
王敏 authored
-
- 09 Oct, 2025 1 commit
-
-
zhuwenwen authored
This reverts merge request !219
-
- 24 Sep, 2025 1 commit
-
-
yangshj1 authored
-
- 16 Sep, 2025 1 commit
-
-
yangshj1 authored
-
- 04 Sep, 2025 2 commits
- 29 Aug, 2025 1 commit
-
-
lizhigong authored
-
- 27 Aug, 2025 1 commit
-
-
lizhigong authored
-
- 25 Aug, 2025 2 commits
- 18 Aug, 2025 1 commit
-
-
lizhigong authored
-
- 16 Aug, 2025 2 commits
- 15 Aug, 2025 1 commit
-
-
lizhigong authored
-
- 11 Aug, 2025 1 commit
-
-
lizhigong authored
-
- 10 Jul, 2025 1 commit
-
-
zhuwenwen authored
-
- 26 May, 2025 1 commit
-
-
lizhg1 authored
-
- 23 May, 2025 1 commit
-
-
lizhigong authored
-