Skip to content

Pull requests: vllm-project/vllm

New pull request New
1,214 Open 14,217 Closed
1,214 Open 14,217 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

[Build] Revert triton_kernels requirements ci/build
#27659 opened Oct 28, 2025 by varun-sundar-rabindranath Loading…
1
[Misc] Make LayerBlockType a Literal instead of Enum ready ONLY add when PR is ready to merge/full CI is needed tpu Related to Google TPUs v1
#27658 opened Oct 28, 2025 by DarkLight1337 Loading…
5 tasks
3
[CI/Build] Move pre-commit only scripts to tools/pre_commit ci/build ready ONLY add when PR is ready to merge/full CI is needed
#27657 opened Oct 28, 2025 by DarkLight1337 Loading…
5 tasks
5
[FLA] Introduce Kimi Delta Attention(KDA) to VLLM
#27654 opened Oct 28, 2025 by zhiyuan1i Loading…
5 tasks
7
[ROCm] [AITER] Add block scaled bpreshuffle gemm rocm Related to AMD ROCm
#27652 opened Oct 28, 2025 by tjtanaa • Draft
5 tasks
[cpu][feat] Enable SWA and hybrid local-global models for CPU backend v1
#27650 opened Oct 28, 2025 by fadara01 Loading…
2 of 5 tasks
9
[KV offload] Offloading connector async scheduling support kv-connector
#27648 opened Oct 28, 2025 by KevinCheung2259 Loading…
5 tasks
4
[Core] Support async-scheduling with kvconnector v1
#27647 opened Oct 28, 2025 by KevinCheung2259 Loading…
5 tasks
1
[MTP] Refactor mtp predictor to avoid d2h operation deepseek Related to DeepSeek models
#27643 opened Oct 28, 2025 by MengqingCao Loading…
5 tasks
3
[Frontend] Add vllm bench sweep to CLI ci/build documentation Improvements or additions to documentation frontend performance Performance-related issues ready ONLY add when PR is ready to merge/full CI is needed
#27639 opened Oct 28, 2025 by DarkLight1337 Loading…
5 tasks
9
feature: support ipv6 in vllm distributed
#27638 opened Oct 28, 2025 by brook-cpp Loading…
5 tasks
5
[Bugfix] vLLM should check Inductor config for compile cache enablement status
#27637 opened Oct 28, 2025 by gmagogsfm Loading…
7
Fix excessive logging noise by reducing the log level of the MinimaxM2ToolParser import success message frontend tool-calling
#27635 opened Oct 28, 2025 by minatoaquaMK2 Loading…
3 of 5 tasks
3
[Core][Env Cache] Enable environment variable cache for Worker as well v1
#27632 opened Oct 28, 2025 by Jialin Loading…
3 of 5 tasks
4
[NIXL][XPU] update name of nixl wheel kv-connector ready ONLY add when PR is ready to merge/full CI is needed
#27631 opened Oct 28, 2025 by zhenwei-intel Loading…
2
[Core][Bookkeeping] Update cu_num_accepted_tokens for all req_index v1
#27629 opened Oct 28, 2025 by Jialin Loading…
3 of 5 tasks
2
Fix MiniMax-M2 rmsnorm precision and remove useless code ready ONLY add when PR is ready to merge/full CI is needed
#27627 opened Oct 28, 2025 by rogeryoungh Loading…
5 tasks
2
Update WSL detection to check for WSL1 instead of "microsoft" as WSL2…
#27625 opened Oct 28, 2025 by HoneyBerries Loading…
3 of 5 tasks
5
[ROCm][Platform] Add MI308X device id in _ROCM_DEVICE_ID_NAME_MAP rocm Related to AMD ROCm
#27623 opened Oct 28, 2025 by sammysun0711 Loading…
3 of 5 tasks
4
[Add] cmdline argument parsing for KV cache offloading modules
#27621 opened Oct 28, 2025 by ApostaC Loading…
6 tasks
1
5
[Model] make the inv_freq in Qwen2_5_VisionRotaryEmbedding device-agnostic qwen Related to Qwen models
#27617 opened Oct 28, 2025 by yangulei Loading…
5
[BUG] Make 'binary' default option for saving torch compile artifacts when using standalone_compile ready ONLY add when PR is ready to merge/full CI is needed
#27616 opened Oct 28, 2025 by ahao-anyscale • Draft
2 of 5 tasks
1
[AsyncScheduling] Make async overlap work with logprobs ready ONLY add when PR is ready to merge/full CI is needed v1
#27615 opened Oct 27, 2025 by njhill Loading…
4
[WIP] Enable async scheduling by default kv-connector ready ONLY add when PR is ready to merge/full CI is needed structured-output suppress-bc-linter tpu Related to Google TPUs v1
#27614 opened Oct 27, 2025 by njhill Loading…
8
add a warmup mode in CUDAGraphMode v1
#27613 opened Oct 27, 2025 by bangshengtang Loading…
10
ProTip! Type g i on any issue or pull request to go back to the issue listing page.