Skip to content

Pull requests: vllm-project/vllm

New pull request New
695 Open 9,039 Closed
695 Open 9,039 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[full_graph] Fix query_start_loc padding v1
#19321 opened Jun 7, 2025 by yinghai Loading…
3 tasks done
9
[v1] Add fp32 support to v1 engine through flex attn ready ONLY add when PR is ready to merge/full CI is needed
#19319 opened Jun 7, 2025 by Isotr0py Loading…
3 tasks done
4
[Bugfix] Fix auto dtype casting for BatchFeature ready ONLY add when PR is ready to merge/full CI is needed
#19316 opened Jun 7, 2025 by Isotr0py Loading…
2 of 3 tasks
1
4
Add H20-3e fused MoE kernel tuning configs for Qwen3-235B-A22B moe qwen Related to Qwen models ready ONLY add when PR is ready to merge/full CI is needed
#19315 opened Jun 7, 2025 by Xu-Wenqing Loading…
3 tasks done
6
[Fix] Remove unused opentelemetry-semantic-conventions-ai dependency ci/build documentation Improvements or additions to documentation
#19313 opened Jun 7, 2025 by conroy-cheers Loading…
3
[Bugfix][V1] Fix memory profile to allow multiple servers to start on the same card needs-rebase v1
#19312 opened Jun 7, 2025 by yeqcharlotte Loading…
7
[Misc]: refactor: ParallelConfig init func
#19310 opened Jun 7, 2025 by googs1025 Loading…
3 tasks
5
[doc] improve ci doc ci/build documentation Improvements or additions to documentation
#19307 opened Jun 7, 2025 by reidliu41 Loading…
3 tasks
5
Use xla flag to improve the quantized model performance ready ONLY add when PR is ready to merge/full CI is needed tpu Related to Google TPUs v1
#19303 opened Jun 6, 2025 by vanbasten23 Loading…
3 tasks done
4
[Misc] Change tests/compile to use VLLM_V1 by default ready ONLY add when PR is ready to merge/full CI is needed
#19302 opened Jun 6, 2025 by zou3519 Loading…
5
Add optional token-level progress bar to LLM.beam_search using tqdm frontend
#19301 opened Jun 6, 2025 by NekoMimiUnagi Loading…
3 tasks done
5
[Misc] Fixes and Optimizations for DeepEP + DeepGEMM combination. ready ONLY add when PR is ready to merge/full CI is needed v1
#19298 opened Jun 6, 2025 by varun-sundar-rabindranath Loading…
15
[CI] Update FlashInfer to 0.2.6 ci/build
#19297 opened Jun 6, 2025 by mgoin Loading…
7
[Quantization] Bump compressed-tensors version ci/build ready ONLY add when PR is ready to merge/full CI is needed
#19295 opened Jun 6, 2025 by kylesayrs Loading…
4
[V1] Add API docs for EncoderCacheManager ready ONLY add when PR is ready to merge/full CI is needed v1
#19294 opened Jun 6, 2025 by russellb Loading…
4
[TPU] support fp8 kv cache quantization tpu Related to Google TPUs v1
#19292 opened Jun 6, 2025 by yaochengji Loading…
7
[Metrics] Compute and log the serving FLOPs documentation Improvements or additions to documentation
#19290 opened Jun 6, 2025 by sysradium Loading…
1
13
[Misc] Add documentation update reminder to PR template ci/build
#19289 opened Jun 6, 2025 by Isotr0py Loading…
1 of 3 tasks
4
[Frontend] Remove unreachable code from llm.py frontend
#19288 opened Jun 6, 2025 by KsuParkhamchuk Loading…
4
[Core] Update error message for Whisper + num-scheduler-steps > 1 ready ONLY add when PR is ready to merge/full CI is needed
#19286 opened Jun 6, 2025 by russellb Loading…
3
[V1][Kernel] Flashinfer HND KV cache layout v1
#19280 opened Jun 6, 2025 by NickLucche Loading…
7
[Frontend] optimize beam_search code
#19267 opened Jun 6, 2025 by zhanggzh Loading…
5
Fix TorchAOConfig skip layers
#19265 opened Jun 6, 2025 by mobicham Loading…
13
Convert kv_transfer_config from dict to KVTransferConfig to fix #19259 frontend
#19262 opened Jun 6, 2025 by maobaolong Loading…
1
4
[New Model]: Support Qwen3 Embedding & Reranker frontend
#19260 opened Jun 6, 2025 by noooop Loading…
2
@DarkLight1337
16
ProTip! Add no:assignee to see everything that’s not assigned.