Commits vllm-project/vllm

[CI] Rebalance multimodal extended generation shards

khluu committed 7 days ago

3eef6435

[CI] Shard long multimodal test groups

khluu committed 7 days ago

f9771fd7

[CI] Add repository-local OTel tracing helpers (#52851)

khluu committed 8 days ago

Verified 504bb8b0

[CPU][Zen] Route Int8 MoE inference through zentorch on AMD (#44834)

ganeshr10 committed 8 days ago

Verified 8f03625b

[Bugfix][KV Offload] Isolate tiering shutdown failures (#52290)

Alex-ai-future committed 8 days ago

Verified c35551f8

[Bugfix][MiniCPM-V] Route video_embeds to the shared vision parser (#54633)

subhashpolisetti committed 8 days ago

Verified 4707679c

[XPU][TEST] Add entrypoints test in Intel GPU CI (#53980)

zxd1997066 committed 8 days ago

Verified 5414b4e6

[Bugfix][Security] Bound embedding densification before to_dense() (#54632)

lzhan011 committed 8 days ago

Verified ce739171

[Bugfix][Frontend] Preserve token offset origins after left text pre-trimming (#54692)

waizuichougou committed 8 days ago

Verified 89056336

[CI][ROCm] Expand weight loading test coverage on AMD and cap its KV cache (#54037)

stefankoncarevic committed 8 days ago

Verified b65af5e3

[Quantization][Refactor][1/N] Adopt `QuantKey` in `QuarkConfig` and methods, relying on `weight_quant_key`, `act_quant_key` for quant method dispatch (#52958)

fxmarty-amd committed 8 days ago

Verified 63988f3c

[Distributed] Support pre-shared ncclUniqueId rendezvous for weight transfer (#53784)

dharak-cohere committed 8 days ago

Verified d0e695a9

[Bugfix][Security] Bound the validation-error response body (#54684)

lzhan011 committed 8 days ago

Verified fa99a6fe

[Core] Release NCCL communicator memory in sleep mode (#51485)

aoshen02 committed 8 days ago

Verified 4ac452ad

[Feature][MM_UUIDs] Allow empty video URLs when using multi-modal UUIDs (#54220)

Jankwi committed 8 days ago

Verified ec32f669

[Bugfix][Frontend] Truncate the assistant tokens mask with the prompt (#54539)

Hotragn committed 8 days ago

Verified ff0c3cb0

[DecodeBenchConnector] Fix HMA cache-group mapping (#54647)

majunze2001 committed 8 days ago

Verified 30dd1a79

[ROCm][CI] Stabilize the sqrt-softplus top-k tie oracle (#54403)

AndreasKaratzas committed 8 days ago

Verified 40b2f620

[Attention][DSA] Enable W4A16 DSA (#51724)

sychen52 committed 8 days ago

Verified 7c5dc571

[ROCm] Keep GLM-5.2 on MRV1 and disable default breakable cudagraph (#53155)

Rohan138 committed 8 days ago

Verified ce2e343b

[ROCm][MoE] Split AITER CK and Triton MXFP4 W4A16 into separate backends (#50622)

afriedri committed 8 days ago

Verified dc9114b2

[Distributed] Add opt-in FlashInfer PCIe IPC all-reduce backend (#53576)

lucifer1004 committed 8 days ago

Verified 446c7694

[1/N][warmup][DSv4] Migrate generic MLA metadata and indexing kernels (#50175)

LopezCastroRoberto committed 8 days ago

Verified e16b5e51

[Bugfix][EC Connector] Fall back when MADV_POPULATE_WRITE is unsupported (#53190)

wentian-byte committed 8 days ago

Verified 188716ac

[Feat] Add request-level preemption count histogram metric (#49984)

linamy85 committed 8 days ago

Verified 923949e6

[Kernel] add Flashinfer cutedsl w4a16 linear (#53014)

IwakuraRein committed 8 days ago

Verified 882ca8d6

[CI][ROCm] Avoid redundant image pulls during smoke validation (#54408)

AndreasKaratzas committed 8 days ago

Verified 907b1a7f

[Core][MRV2] Freeze gc during V2 CG capture; skip per-descriptor cleanup (#54646)

njhill committed 8 days ago

Verified c28feab9

[Rust Frontend] Migrate to new tekken crate (#53056)

jorge-menjivar committed 8 days ago

Verified 225aec48

[CI] Broaden tool-calling issue auto-labeling (#54650)

sfeng33 committed 8 days ago

Verified 45aed9b0