Commits vllm-project/vllm

Merge remote-tracking branch 'origin/main' into lwilkinson/fix-glm-5-mtp-more-then-1

LucasWilkinson committed 48 days ago

9fc312d7

[Attention] FA4 integration (#32974)

LucasWilkinson committed 48 days ago

Verified 8b5014d3

Revert "[Bugfix] Disable TRTLLM attention with KV transfer enabled (#33192)" (#34832)

ZhanqiuHu committed 48 days ago

Verified 57a96e26

[torch.compile] Undo the fast_moe_cold_start hack in torch>=2.11 (#35475)

zou3519 committed 48 days ago

Verified e82fbeec

[Bugfix] Fix dtype mismatch in RMSNormGated.forward_native() during torch.compile (#35256)

haosdent committed 48 days ago

Verified 62904708

[Model Runner V2] Use block table apis for capture inputs (#35671)

WoosukKwon committed 49 days ago

Verified 72f4d162

fix(mxfp4): return is_monolithic=False when LoRA is enabled for Triton backend (#35382)

yoonsnowdev committed 49 days ago

Verified 5a435507

[MISC] Fixing a null reference by removing parallel_utils from mypy EXCLUDE (#35630)

taneem-ibrahim committed 49 days ago

Verified 59d7af9c

[Mamba1] - Kernel Level Chunk Alignment for Prefix Caching (#34798)

Josephasafg committed 49 days ago

Verified bbf81f9a

[Model Runner V2] Minor refactoring for EncoderRunner (#35628)

WoosukKwon committed 49 days ago

Verified da543d1a

[AMD][CI] Support Triton attention with ExampleConnector (#34931)

rjrock committed 49 days ago

Verified 87d319c5

Fix typo: implictly -> implicitly in isaac.py docstring (#35646)

lin-shh committed 49 days ago

Verified a9ec392c

[Bugfix][Model] Fix Qwen3.5/Qwen3Next ignoring --dtype flag on older GPUs (#35617)

lailoo committed 49 days ago

Verified afd089f2

Add TMA support to fused_moe_lora kernel (#32195)

gnovack committed 49 days ago

Verified 3ecd0bf9

[Model Runner V2] Add ModelStateInterface [4/N] (#35621)

WoosukKwon committed 49 days ago

Verified e3eb146f

[Bugfix] Fix Anthropic API base64 image handling in Messages endpoint (#35557)

voipmonitor committed 49 days ago

Verified 95a395db

[Chore] Cleanup BNB utilization dead code (#35620)

Isotr0py committed 50 days ago

Verified e94b263b

[Deprecation] Deprecate code in 0.17 as scheduled (#35441)

yewentao256 committed 50 days ago

Verified e113a301

[Benchmark] Avoid unnecessary video download in MMVU (#35618)

DarkLight1337 committed 50 days ago

Verified 1dafb29f

[Fix] Avoid sending image input to other PP ranks (#35405)

emricksini-h committed 50 days ago

Verified 49b9ae32

Fix Qwen3_5MTP packed_modules_mapping for gate_up_proj (#35581)

cwazai committed 50 days ago

Verified 63d7972f

custom dataset img support base64 (#35280)

flutist committed 50 days ago

Verified c68e69f1

[Feat] Add CUDA torch fallbacks for fp8_mqa_logits/fp8_paged_mqa_logits_torch function (#35271)

chaunceyjiang committed 50 days ago

Verified 7e08c22b

add io_process_plugin for sparse embedding (#34214)

staugust committed 50 days ago

Verified 8e75d885

[Feature]Supports Anthropic Thinking Block (#33671)

mariohong128 committed 50 days ago

Verified 0892d1ab

Add padding support to wvSplitK solution for skinny GEMMs (#33762)

amd-hhashemi committed 50 days ago

Verified 7600642e

[ROCm][CI] Parametrize vision score tests across attention backends with per-backend tolerances (#35571)

AndreasKaratzas committed 50 days ago

Verified 1e69c048

[Benchmark] Improve UX of sweep scripts (#35600)

DarkLight1337 committed 50 days ago

Verified 4292e3b8

[Benchmark] Rename SLA Finder to Workload Explorer (#35586)

DarkLight1337 committed 50 days ago

Verified 24d6ea8a

[Misc] Change logging level from info to debug for tool parser import (#35575)

chaunceyjiang committed 50 days ago

Verified 57c86c07