vllm-project/vllm

Pull Requests Commits

Merge branch 'main' into fix-aiter-mixtral

tjtanaa committed 135 days ago

Verified 96088d9b

Add Fused MoE Triton kernels for GLM-4.5-Air, GLM-4.5v, GLM-4.6v on 2x RTX Pro 6000 (#31407)

mratsim committed 135 days ago

Verified b9793e6a

[Model] Add tuned triton fused_moe configs for Qwen3Moe on B200 (#31448)

Jzz1943 committed 135 days ago

Verified 0b6b7010

[BugFix] Re-fix async multimodal cpu tensor race condition (#31373)

njhill committed 136 days ago

Verified 094fcce2

[ROCm] Migrate xgrammar to upstream release (#31327)

AndreasKaratzas committed 136 days ago

Verified 573dd0e6

[ROCm][CI] Add TorchCodec source build for transcription tests (#31323)

AndreasKaratzas committed 136 days ago

Verified f7036886

[ROCm][CI] Added perceptron lib in requirements for isaac multi-modal test (#31441)

AndreasKaratzas committed 136 days ago

Verified 96142f20

[BugFix] register quant scale tensors as buffer (#31395)

BoyuanFeng committed 136 days ago

Verified 62def07d

add tip for VLLM_USE_PRECOMPILED arg to reduce docker build time (#31385)

yitingdc committed 136 days ago

Verified b326598e

[MoE Refactor][10/N] Cleanup Fp8 Process Weights After Loading (#31169)

robertgshaw2-redhat committed 136 days ago

Verified 727c41f3

[BugFix] Fix cache issue in compilation_config (#31376)

BoyuanFeng committed 136 days ago

Verified 2f12cd32

[Chore]: Remove HF format Phi4-MM examples (#31405)

Isotr0py committed 137 days ago

Verified 40a87562

[CI/Build] Ignore max transformers version for more common tests (#31401)

Isotr0py committed 137 days ago

Verified 3d024985

Fix/get raw stream patch #30905 (#30912)

baonudesifeizhai committed 137 days ago

Verified 8711b216

[Core][Hybrid allocator + connector] Support hybrid allocator + kv cache connector (#30166)

ivanium committed 137 days ago

Verified 52bf0665

[XPU][CI]skip test_preprocess_error_handling due to fork/spawn issue (#31381)

jikunshang committed 137 days ago

Verified 5326c898

CustomOp: Unify aiter impl into GroupedTopk (#31221)

xinyu-intel committed 137 days ago

Verified 87f1b8ca

[Docs] Add profiler user docs for http request (#31370)

lengrongfu committed 137 days ago

Verified 887e900b

[Mistral common] Ensure all functions are imported from the top & only use public methods (#31138)

patrickvonplaten committed 138 days ago

Verified 48e74497

[Core] Initialize LoRA support for tower and connector in multi-modal models (#26674)

jeejeelee committed 138 days ago

Verified ce1eafd1

[Docs] Fix some snippets (#31378)

hmellor committed 138 days ago

Verified 0b544e64

[Misc] Fix Qwen2-MoE shared_expert_gate (#31339)

jeejeelee committed 138 days ago

Verified c3666f56

[CI] Fix flaky vision beam search test with flexible semantic validation (#31324)

AndreasKaratzas committed 138 days ago

Verified c79dbfa9

Support LoRA and GPTQModel for PLaMo 2/3 (#31322)

Alnusjaponica committed 138 days ago

Verified 9ee05cbe

[benchmark] use model card root instead of id (#31329)

andyxning committed 138 days ago

Verified 3b8f31b3

[CI/Build] Ignore max transformers version skipping for initialization tests (#30619)

Isotr0py committed 138 days ago

Verified 2cd94259

Feature/isaac 0.1 (#28367)

oscardev256 committed 138 days ago

Verified b7165d53

[BugFix] Fix async scheduling + reasoning with struct output (#31332)

njhill committed 138 days ago

Verified 81786c87

[Hybrid] Mamba2 prefix cache blocks freeing for running requests (#28047)

s3woz committed 138 days ago

Verified f1531d9f

[Model][Ernie4.5-VL] Support video metadata for timestamp rendering (#31274)

Tiiiktak committed 138 days ago

Verified 2d6001f4

Older