Skip to content

Pull requests: vllm-project/vllm

New pull request New
3,780 Open 27,680 Closed
3,780 Open 27,680 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

[Bugfix][LoRA] Add embedding_modules for Qwen3.5 CausalLM bug Something isn't working qwen Related to Qwen models
#48850 opened Jul 16, 2026 by Agoni-02 Loading…
1
Fix: Restore data_parallel_size > 1 for use_sequence_parallel_moe
#48849 opened Jul 16, 2026 by passtoor-agi Loading…
3 of 4 tasks
1
[ROCm][CI] Loosen block-FP8 fused MoE test tolerance for large-K shapes rocm Related to AMD ROCm
#48847 opened Jul 16, 2026 by stefankoncarevic Contributor Loading…
4 tasks
[Bugfix][Tool Parser] MiniMax M2: preserve whitespace in parameter values bug Something isn't working tool-calling
#48846 opened Jul 16, 2026 by mosya415 Loading…
1
[ROCm][CI] Fix AITER MLA fp8 decode metadata regression test rocm Related to AMD ROCm
#48845 opened Jul 16, 2026 by stefankoncarevic Contributor Loading…
4 tasks
[Doc] Update CPU image references to Docker Hub cpu Related to CPU backends documentation Improvements or additions to documentation
#48844 opened Jul 16, 2026 by lizzzcai Contributor Loading…
1
[BugFix] Set graph_pool_id before FULL CUDA graph capture in ModelRunner V2 bug Something isn't working nvidia v1
#48843 opened Jul 16, 2026 by ilmarkov Contributor Loading…
3 of 5 tasks
1
[Bugfix] Fix int4_per_token_head for SWA models and gemma 4 bug Something isn't working v1
#48842 opened Jul 16, 2026 by VikramRangarajan Loading…
3 of 4 tasks
1
2
[ROCm] [Model] Enable TML inkling ci/build needs-rebase new-model Requests to new models nvidia performance Performance-related issues rocm Related to AMD ROCm rust tool-calling v1
#48841 opened Jul 16, 2026 by tjtanaa Member • Draft
4 tasks
1
[docs] preserve page path in stable-docs announcement link build-docs documentation Improvements or additions to documentation verified Run pre-commit for new contributors without triggering other tests
#48839 opened Jul 16, 2026 by sagearc Contributor Loading…
@hmellor
2
[V1][Core] Add public build_full_block_hash_chain helper for offline prefix-cache analysis v1
#48838 opened Jul 16, 2026 by raravind007 • Draft
1
[ZenCPU] Changes with respect to cpu model tests for ZenDNN ci/build cpu Related to CPU backends
#48836 opened Jul 16, 2026 by Chinmay-Kulkarni-AMD Contributor • Draft
4 tasks
[Bugfix][Multimodal] Fix M-RoPE media mapping at chunked-prefill boundaries bug Something isn't working multi-modality Related to multi-modality (#4194)
#48835 opened Jul 16, 2026 by parkilwoo Loading…
4 tasks done
1
@Isotr0py
1
[ZenCPU] Changes with respect to cpu platform detection test for ZenDNN ci/build cpu Related to CPU backends
#48834 opened Jul 16, 2026 by Chinmay-Kulkarni-AMD Contributor • Draft
4 tasks
[Frontend]Flatten beam-search beams with itertools.chain instead of sum frontend ready ONLY add when PR is ready to merge/full CI is needed
#48829 opened Jul 16, 2026 by wangxingda Loading…
3
[XPU] allow forcing flash attn for mm_prefix intel-gpu Related to Intel GPU
#48828 opened Jul 16, 2026 by zhenwei-intel Contributor Loading…
4 tasks
1
Perf/h20 moe config e256 n512
#48825 opened Jul 16, 2026 by zzt93 Contributor Loading…
3 of 4 tasks
fix(security): validate bad_words token IDs against model vocab size rust v1
#48824 opened Jul 16, 2026 by jperezdealgaba Contributor Loading…
[Model] Add breakable CUDA graph support for Inkling [2/N] nvidia ready ONLY add when PR is ready to merge/full CI is needed v1
#48822 opened Jul 16, 2026 by WoosukKwon Collaborator Loading…
fix: defer optional transformers model imports
#48820 opened Jul 16, 2026 by host452b Loading…
1
1
[ZenCPU] Changes for relevant cpu tests and image build for ZenDNN ci/build cpu Related to CPU backends
#48818 opened Jul 16, 2026 by Chinmay-Kulkarni-AMD Contributor • Draft
4 tasks
[Bugfix]Fix hybrid model sliding window RuntimeError bug Something isn't working v1
#48817 opened Jul 16, 2026 by Yejing-Lai Contributor Loading…
Fix GPTQ quantized Qwen3.5 MTP weight loading with spec decode qwen Related to Qwen models
#48816 opened Jul 16, 2026 by vllmellm Contributor Loading…
2 of 4 tasks
Fix MTP Mamba align prefix cache retention v1
#48815 opened Jul 16, 2026 by YuYue1208 • Draft
1
[Kernel] Remove unused N/B constexpr params from pack/unpack seq kernels v1
#48813 opened Jul 16, 2026 by aryan331277 Loading…
1
ProTip! Follow long discussions with comments:>50.