vllm-project/vllm

Pull Requests Commits

WoosukKwon committed 78 days ago

22bf5c50

WoosukKwon committed 78 days ago

3a899074

WoosukKwon committed 78 days ago

fbc2cc82

[Benchmark] Support Infinity API (#26641)

DarkLight1337 committed 78 days ago

Verified 5be7ca1b

[Bugfix] Fix qwen-moe packed_modules_mapping (#26634)

jeejeelee committed 78 days ago

Verified f0a30a06

[Bugfix][Qwen3VL] fix deepstack in qwen3vl (#26626)

JJJYmmm committed 78 days ago

Verified 9d6cff3e

[compile] Add patched_fused_scaled_matmul_reduce_scatter (#26604)

angelayi committed 78 days ago

Verified a25f2ade

[Refactor]Reduce duplicate code in serving_chat (#26627)

chaunceyjiang committed 78 days ago

Verified d0bed837

[CPU] fix the issue when the node is '-' cause json decode error. (#26562)

muzian666 committed 78 days ago

Verified f7ee6986

Add EAGLE-3 Speculative Decoding Support for Qwen3 MoE (#26485)

rahul-tuli committed 78 days ago

Verified d2a71530

fix(nix): Allow local oneDNN path to fix vLLM CPU build failure (#26401)

ihb2032 committed 78 days ago

Verified 086609de

[Refactor]: Use M-RoPE interface directly while defining model class instead of maintaining model specific M-RoPE implementation in mrope.py (#24172)

divyanshsinghvi committed 78 days ago

Verified 727144be

[Bugfix][Multi Modal] Fix incorrect Molmo image processing (#26563)

sangho-vision committed 78 days ago

Verified 55392bc8

[MM] Move Qwen3Omni MRoPE impl to model file (#26608)

ywang96 committed 79 days ago

Verified ddaff293

[XPU] Upgrade NIXL to remove CUDA dependency (#26570)

zhenwei-intel committed 79 days ago

Verified 27ed39a3

[CI/Build] Fix ppc64le CPU build and tests (#22443)

npanpaliya committed 79 days ago

Verified 8f8474fb

[Frontend] Improve the performance of `is_reasoning_end` (#25735)

chaunceyjiang committed 79 days ago

Verified be067861

[BugFix] Make penalties and bad_words work with async scheduling (#26467)

njhill committed 79 days ago

Verified 5bc26c43

AOT Compilation for torch.compile (Bundled) (#24274)

zhxchen17 committed 79 days ago

Verified eef921f4

Cache the environment variable check for batch invariance (#26510)

bwasti committed 79 days ago

Verified e317414c

[BugFix] Fix async scheduling + request preemption (#26385)

njhill committed 79 days ago

Verified 949cb017

[BUG] Qwen3-next MTP. Fix attn metadata build bug (#26564)

vadiklyutiy committed 79 days ago

Verified e94cfd51

Fix some typing issues found by `mypy==1.18.2` (#26596)

hmellor committed 79 days ago

Verified 7c12763b

Update CUDA architecture list in build pipeline for 12.9.1 wheels (#26592)

wseaton committed 79 days ago

Verified 3b780a4b

Update `pre-commit` hook versions (#26591)

hmellor committed 79 days ago

Verified 30f78af1

Add Qwen3-Omni moe thinker (#25550)

wangxiongts committed 79 days ago

Verified 19a9b169

[Transform] [Quantization] Add QuTLASS support to vLLM (#24440)

LopezCastroRoberto committed 79 days ago

Verified 96ad65b7

[Model] Add FlexOlmo model implementation (#24923)

2015aroras committed 79 days ago

Verified 8d2b8c0f

[Model][Qwen3VL] Compute `cu_seqlens` on CPU to remove (#26496)

lgeiger committed 79 days ago

Verified b2155ed3

[Bugfix] fixed top_logprobs: -1 does not appear to work as intended (#26470)

chaunceyjiang committed 79 days ago

Verified 910abdbd

Older