vllm-project/vllm

Pull Requests Commits

[bugfix] Qwen3-VL fix video incorrect timestamp calculations while do_sample_frames=True (#27104)

wulipc committed 52 days ago

Verified 4c91a28e

[Data-parallel] Allow DP>1 for world_size > num_gpus on node (8) (#26367)

patrickvonplaten committed 52 days ago

Verified b038d9c4

[CI] Nixl integration tests (#27010)

NickLucche committed 52 days ago

Verified 2ba60ec7

[torch.compile] Enable attention and allreduce fusion without custom ops enabled (#24604)

ProExpertProg committed 52 days ago

Verified bd7157a0

Fix incorrect docstring for stop_profile() method (#27101)

hyongtao-code committed 52 days ago

Verified be429d0c

[Harware][AMD][Model] Triton MoE tuning configs for GLM-4.5 for MI350 and MI355 (#25586)

rkarhila-amd committed 52 days ago

Verified c253745e

[Model]Improve Qwen3VLMoeForConditionalGeneration packed_modules_mapping (#27096)

jeejeelee committed 52 days ago

Verified daec4d26

[Docs] Replace `rst` style double-backtick with `md` single-backtick (#27091)

hmellor committed 52 days ago

Verified 6c9fdbf7

[Docs] Replace all explicit anchors with real links (#27087)

hmellor committed 52 days ago

Verified 483ea646

[VLM][Refactor] Remove useless func `get_input_positions` in `MRotaryEmbedding` (#27088)

MengqingCao committed 52 days ago

Verified e20eba75

Update troubleshooting.md and remind VLLM_TRACE_FUNCTION usage (#27069)

Prowindy committed 52 days ago

Verified bbc1b296

[CI] fix docs build failed (#27082)

chaunceyjiang committed 52 days ago

Verified acb1bfa6

[Kernel][Performance] Fuse float cast and renormalize to topk softmax kernel (#26717)

izhuhaoran committed 52 days ago

Verified 75c7ad99

[CI/Build] Update compressed tensor test path to fix CPU CI (#27068)

bigPYJ1151 committed 52 days ago

Verified 5550ff9c

[Model] Add support for LightOnOCR (#26916)

staghado committed 52 days ago

Verified 3aeb19a3

[Model] Always use Transformers backend for PaliGemma and Gemma3-MM (#26715)

DarkLight1337 committed 52 days ago

Verified 8c017b34

[CI/Build] Update Llama4 eval yaml (#27070)

zhewenl committed 52 days ago

Verified 9c2c2287

[Kernel] Lazy import FlashInfer (#26977)

jeejeelee committed 52 days ago

Verified fec2b341

[Bugfix] Fix ReplicatedLinearWithLoRA (#27065)

jeejeelee committed 52 days ago

Verified 87bc0c49

[Core] Change `execute_model_with_error_logging()` to be a ctx manager (#27060)

njhill committed 53 days ago

Verified fe3b9372

[Bugfix][Qwen] fixes the weights dtype in qwen3_next: it is actually a bfloat16 (#27030)

sighingnow committed 53 days ago

Verified bde9e227

disable graph partition in custom op (#26952)

BoyuanFeng committed 53 days ago

Verified 08405609

[Perf] Exploit out-of-band buffers in shm_broadcast (#26961)

njhill committed 53 days ago

Verified ab81379e

[Docs] Reduce custom syntax used in docs (#27009)

hmellor committed 53 days ago

Verified 4ffd6e89

vllm bench serve shows num of failed requests (#26478)

tomasruizt committed 53 days ago

Verified 965c5f49

Remove unused imports (#26972)

lgeiger committed 53 days ago

Verified 4d055ef4

[torch.compile] fix simple inductor graph partition test (#27050)

BoyuanFeng committed 53 days ago

Verified 17c540a9

[Chore] Separate out `vllm.utils.importlib` (#27022)

DarkLight1337 committed 53 days ago

Verified 4d4d6bad

[torch.compile] Passing only necessary compilation config to inductor pass config (#27041)

luccafong committed 53 days ago

Verified 11ae016b

[NVIDIA] [Perf] Update to leverage flashinfer trtllm FP4 MOE throughput kernel (#26714)

jiahanc committed 53 days ago

Verified 41d30719

Newer Older