Skip to content

Actions: hongxiayang/vllm

Actions

pre-commit

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
pre-commit.yml
66 workflow runs
66 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
[BugFix] Fix prompt_embeds for multimodal models (#45383) pre-commit #84: Commit c621af1 pushed by hongxiayang
4m 7s main
main
4m 7s
[Perf] SM90 cutlass fp8 mm supports odd M by swap_ab, 180~290% kernel… pre-commit #83: Commit 71b961d pushed by hongxiayang
4m 4s main
main
4m 4s
[ROCm][DSV4][Perf] Fuse inverse-RoPE and cache bf16 wo_a in o-project… pre-commit #82: Commit badddd2 pushed by hongxiayang
5m 33s main
main
5m 33s
[Bugfix][CI] Add missing import of pad_nvfp4_activation_for_cutlass i… pre-commit #81: Commit 2a43b40 pushed by hongxiayang
4m 34s main
main
4m 34s
Revert "[torch.compile] Add patch for fullgraph compilation" (#42686)… pre-commit #80: Commit 1ac10f1 pushed by hongxiayang
4m 34s main
main
4m 34s
[XPU] fix weight scale shape (#42725) pre-commit #79: Commit 0fa8884 pushed by hongxiayang
5m 35s main
main
5m 35s
fix(openai): tolerate empty content in forced tool choice (#40148) pre-commit #78: Commit 6467213 pushed by hongxiayang
3m 54s main
main
3m 54s
[FEATURE] Add EagleMistralForCausalLM (#41024) pre-commit #77: Commit e9f8f31 pushed by hongxiayang
4m 3s main
main
4m 3s
[NVFP4] NVFP4 MOE emulation fallback for H100/MI300/MI350, standardiz… pre-commit #76: Commit d622e27 pushed by hongxiayang
4m 14s main
main
4m 14s
[ROCm] Hotfix: guard MLA dual RMS norm fusion against older AITer ver… pre-commit #75: Commit 21b086d pushed by hongxiayang
3m 53s main
main
3m 53s
[Attention] TurboQuant: remove redundant random signs, add prior art … pre-commit #74: Commit ed0622e pushed by hongxiayang
5m 24s main
main
5m 24s
Triton MLA perf fixes (#33529) pre-commit #73: Commit d9408ff pushed by hongxiayang
4m 35s main
main
4m 35s
enable skipping of SW attention layers when using FP8 KV cache (#33695) pre-commit #72: Commit 98e7f22 pushed by hongxiayang
4m 38s main
main
4m 38s
elastic_ep: Fix issues with repeated scale up/down cycles (#37131) pre-commit #71: Commit c57d38d pushed by hongxiayang
3m 4s main
main
3m 4s
[UltraVox] Fix output type (#37224) pre-commit #70: Commit 2ff0ad9 pushed by hongxiayang
3m 10s main
main
3m 10s
[Bugfix] fix paddleocr crash on some image shape (#36959) pre-commit #69: Commit 4508532 pushed by hongxiayang
4m 17s main
main
4m 17s
[LMCache MP Patch]: Race Condition + Duplicated Block Ids (#35831) pre-commit #68: Commit 5d6aae4 pushed by hongxiayang
5m 7s main
main
5m 7s
[Docs] Add breadcrumbs for better UX (#35749) pre-commit #67: Commit 7e9149d pushed by hongxiayang
4m 9s main
main
4m 9s
[MISC] Fixing a null reference by removing parallel_utils from mypy E… pre-commit #66: Commit 59d7af9 pushed by hongxiayang
4m 5s main
main
4m 5s
[Bugfix] Fix DCP + FA3 crash due to missing num_splits in _forward_wi… pre-commit #65: Commit 6d4f9d3 pushed by hongxiayang
3m 39s main
main
3m 39s
Remove padding_index from models that don't use it for better Trans… pre-commit #64: Commit c38b8d5 pushed by hongxiayang
4m 29s main
main
4m 29s
[Bugfix][MTP][Sparse MLA] Allow sparse MLA with MTP to run with FULL … pre-commit #63: Commit dc5fa77 pushed by hongxiayang
3m 56s main
main
3m 56s
[NemotronH] Do not force router to run in fp32 (#34582) pre-commit #62: Commit 3b30e61 pushed by hongxiayang
4m 50s main
main
4m 50s
[Misc] Introduce ec_both role EC (encoder cache) connector (#34182) pre-commit #61: Commit 33bcd3d pushed by hongxiayang
3m 43s main
main
3m 43s
[Feat][RL] Pause and Resume with keep requests for single engine (#32… pre-commit #60: Commit 89a385d pushed by hongxiayang
4m 59s main
main
4m 59s