Skip to content

Pull requests: vllm-project/vllm

New pull request New
927 Open 10,946 Closed
927 Open 10,946 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

[Core][BugFix] Fix thread safety issue in RequestOutputCollector ready ONLY add when PR is ready to merge/full CI is needed v1
#22576 opened Aug 9, 2025 by 22quinn Loading…
3 of 4 tasks
3
[Misc] Replace flaky image urls in pixtral test multi-modality Related to multi-modality (#4194)
#22574 opened Aug 9, 2025 by Isotr0py Loading…
1 of 4 tasks
6
optimize: improve scheduler policy lookup performance v1
#22573 opened Aug 9, 2025 by skyloevil Loading…
4
[Core] Use individual MM items in P0/P1 cache and model runner multi-modality Related to multi-modality (#4194) tpu Related to Google TPUs v1
#22570 opened Aug 9, 2025 by DarkLight1337 Loading…
1 of 4 tasks
11
[Misc] code clean duplicate set_current_vllm_config in _set_vllm_config
#22566 opened Aug 9, 2025 by andyxning Loading…
4 tasks
2
Frontend: Adding LM Format Enforcer support to V1 engine ci/build structured-output v1
#22564 opened Aug 9, 2025 by noamgat Loading…
4 tasks
2
[BugFix] EAGLE Load Bias From Config llama Related to Llama models speculative-decoding
#22558 opened Aug 9, 2025 by MMuzzammil1 Loading…
2
[gpt-oss] Add test for response API + harmony (but skipped)
#22554 opened Aug 9, 2025 by heheda12345 Loading…
3 of 4 tasks
4
[Misc] fail fast when exception is raised in in_the_same_node_as
#22553 opened Aug 8, 2025 by andyxning Loading…
4 tasks
3
[V1][Hybrid][Backend] Allow different data types in Hybrid Cache Manager needs-rebase v1
#22544 opened Aug 8, 2025 by bringlein • Draft
4 tasks
6
[Bugfix][V1] Fix Finished Request Handling in Async Scheduling bug Something isn't working v1
#22543 opened Aug 8, 2025 by leo-cf-tian Loading…
3
[WIP, Do not review] DI v1
#22542 opened Aug 8, 2025 by yarongmu-google Loading…
4 tasks
4
add tg-mxfp4-moe-test
#22540 opened Aug 8, 2025 by IwakuraRein Loading…
3 tasks done
4
[Kernel] Add cuda kernel for gpt_oss activation
#22538 opened Aug 8, 2025 by jeejeelee • Draft
4 tasks
8
New moe quant config ci/build deepseek Related to DeepSeek models documentation Improvements or additions to documentation needs-rebase rocm Related to AMD ROCm
#22537 opened Aug 8, 2025 by bnellnm • Draft
10
Fix torch version check for mxfp4
#22535 opened Aug 8, 2025 by zifeitong Loading…
4 tasks
4
Fix(benchmarks): allow multiple mm contents in OpenAI Chat Completion Benchmarks performance Performance-related issues
#22534 opened Aug 8, 2025 by h-brenoskuk Loading…
8
Improve fast_topk function with type hints and documentation
#22530 opened Aug 8, 2025 by skyloevil Loading…
7
Remove redundant row_indices unsqueeze operation in MiniCPMO
#22528 opened Aug 8, 2025 by skyloevil Loading…
3
Quantization: support FP4 quantized models on AMD CDNA2/CDNA3 GPUs ci/build rocm Related to AMD ROCm
#22527 opened Aug 8, 2025 by fengli1702 Loading…
4 tasks done
4
[Fix] fix offline env use local mode path
#22526 opened Aug 8, 2025 by lengrongfu Loading…
1 of 4 tasks
4
[WIP] [Bench] Add Triton NVFP4 GEMM performance Performance-related issues
#22523 opened Aug 8, 2025 by phuhung273 • Draft
2 of 4 tasks
4
[ROCm][AITER] Support AITER Rope ops in RotaryEmbedding Module. rocm Related to AMD ROCm
#22521 opened Aug 8, 2025 by vllmellm Loading…
3 of 4 tasks
4
code clean when clean shm
#22516 opened Aug 8, 2025 by andyxning Loading…
4 tasks
4
[WIP][Model] Add Ernie4.5 VL Model Support documentation Improvements or additions to documentation new-model Requests to new models
#22514 opened Aug 8, 2025 by CSWYF3634076 • Draft
14
ProTip! Adding no:label will show everything without a label.