Skip to content

Actions: vllm-project/vllm

Actions

New PR Bot

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
new_pr_bot.yml
2,500+ workflow runs
2,500+ 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
Dynamic-fork scheduling, Medusa/MTP spec decode, and InternVL resize for HPD-Parsing (based on v0.17.1) New PR Bot #8675: Pull request #48715 opened by wwwjjjj
9s
9s
Synchronize distributed FlashInfer autotuning and persist cache New PR Bot #8674: Pull request #48714 opened by jiahanc
9s
9s
[debug] 添加了eager 模式下的 权重hook New PR Bot #8673: Pull request #48713 opened by tonyh168
7s
7s
[Bugfix][ROCm] Only run FP8 AITER MLA prefill when using FP8 KV New PR Bot #8672: Pull request #48712 opened by simondanielsson
8s
8s
[Bugfix] Fix GLM5 config New PR Bot #8671: Pull request #48711 opened by jeejeelee
7s
7s
[BugFix] Fix Jamba tool parser deleting all occurrences of the streamed prefix on tool transition New PR Bot #8670: Pull request #48709 opened by chuenchen309
7s
7s
[ROCm][Bugfix] Return empty UUID string when amdsmi uuid query fails New PR Bot #8669: Pull request #48708 opened by ChuanLi1101
8s
8s
[Bugfix][Parser] Fix streaming tool-call args truncated by schema coercion New PR Bot #8668: Pull request #48706 opened by mahadrehmann
7s
7s
[BugFix] Fix FunctionGemma streaming tool-call arguments corrupted once a 2nd key arrives New PR Bot #8667: Pull request #48705 opened by chuenchen309
7s
7s
[Feature] Support prompt cache retention policies for tiered KV offloading New PR Bot #8666: Pull request #48704 opened by chaunceyjiang
7s
7s
[XPU] [UT] [CI] add xpu config to run gpt-oss accuracy in ut and ci New PR Bot #8665: Pull request #48703 opened by zufangzhu
6s
6s
[Bugfix][Quantization] Reduce NVFP4 MoE load memory under sleep mode New PR Bot #8664: Pull request #48701 opened by hubunt
9s
9s
[Bugfix]Fix transformer backend failed: AttributeError: 'Parameter' object has no attribute 'weight_loader' New PR Bot #8663: Pull request #48699 opened by Yejing-Lai
7s
7s
[Bugfix] Scale FlashInfer b12x capacity for data parallelism New PR Bot #8662: Pull request #48698 opened by dumko2001
8s
8s
[Bugfix] Reject DeepSeek V4 FP4 MoE tensor parallelism New PR Bot #8661: Pull request #48697 opened by dumko2001
10s
10s
[Bugfix] Pad misaligned MXFP4 MoE shards for Marlin New PR Bot #8660: Pull request #48696 opened by dumko2001
9s
9s
[Bugfix] Gate FlashInfer FP4 MoE on kernel availability New PR Bot #8659: Pull request #48695 opened by dumko2001
7s
7s
[Bugfix] Avoid FlashAttention 3 with batch invariance on Hopper New PR Bot #8658: Pull request #48694 opened by dumko2001
8s
8s
[Bugfix] Make Gemma 4 suppress-token masking CUDA-graph safe New PR Bot #8657: Pull request #48693 opened by dumko2001
10s
10s
[MRV2][Spec Decode] Adaptive Speculative Decoding - Initial Support New PR Bot #8656: Pull request #48692 opened by benchislett
18m 16s
18m 16s
[Spec Decode] Add support for num draft layers > 1 for Eagle New PR Bot #8655: Pull request #48691 opened by ekagra-ranjan
28m 17s
28m 17s
[Bugfix] MiniMax M3 reasoning parser: treat non-leading <mm:think> as content New PR Bot #8654: Pull request #48690 opened by harjothkhara
26m 40s
26m 40s
[ROCm][Bugfix] Enable the fp32 head_dtype torch.mm fast path on ROCm New PR Bot #8653: Pull request #48688 opened by wjabbour
7s
7s
[chore] code refactor New PR Bot #8652: Pull request #48687 opened by andyxning
16s
16s
[Model] Add minimal native RWKV7 serving support New PR Bot #8651: Pull request #48686 opened by shiroko98
4m 59s
4m 59s