vllm-project/vllm

Pull Requests Commits

[BugFix] Fix DeepGEMM warmup, no m.weight_scale_inv (#25206)

LucasWilkinson committed 81 days ago

Verified 9fac6aa3

[KV offload][1b/N] rename offloading to kv_offload (#25191)

orozery committed 81 days ago

Verified a53ad626

[V0 Deprecation] Remove unused async_timeout.py (#25190)

WoosukKwon committed 81 days ago

Verified 1c3dad22

[Bug] Fix torch Compilation Cache Hit Error (#25093)

yewentao256 committed 81 days ago

Verified d2a30a2d

[Bug] Fix `returned_lse` not Defined issue (#25106)

yewentao256 committed 81 days ago

Verified 75fb112d

[feat]: Create interface for model-specific M-RoPE (#24194)

AzizCode92 committed 81 days ago

Verified 38db529f

[fix]: remove data type hardcoding from gptoss model implementation (#23807)

nikhil-arm committed 81 days ago

Verified 064cac7b

[V0 Deprecation] Remove AsyncLLMEngine (#25025)

WoosukKwon committed 81 days ago

Verified e19bce40

[KV offload][1/N] Introduce an offloading component (#19848)

orozery committed 81 days ago

Verified 505805b6

[ROCm][AITER][Bugfix] Switch AITER to use PIECEWISE_AND_FULL compilation (#25104)

Rohan138 committed 81 days ago

Verified bbdc0f23

[ROCm][CI/Build] Use ROCm7.0 as the base (#25178)

gshtras committed 81 days ago

Verified dc340593

[spec decode] Fix MTP inference path for MiMo-7B model (#25136)

zixi-qi committed 81 days ago

Verified c4cb0af9

[Misc] Add codeowner for Transformers backend (#25180)

hmellor committed 81 days ago

Verified 1c3b1634

Enable Allgather/ReduceScatter backend for NaiveAllToAll (#23964)

wenscarl committed 81 days ago

Verified 2ea50e97

[Docs] Fix warnings in mkdocs build (continued) (#25163)

Zerohertz committed 81 days ago

Verified b419937c

[New Model] Support BertForTokenClassification / Named Entity Recognition (NER) task (#24872)

noooop committed 81 days ago

Verified 5f696c33

feat(api): Return 503 on /health when engine is dead (#24897)

dongbo910220 committed 81 days ago

Verified 67244c86

[PERF] Add `conv1d` metadata to GDN attn (#25105)

vadiklyutiy committed 81 days ago

Verified 072d7e53

[Kernel] Decouple Tile Size from Block Size in Triton Unified Attention Kernel (#21197)

jvlunteren committed 81 days ago

Verified 01a583fe

[Misc] Add kv-connector label (#25156)

NickLucche committed 81 days ago

Verified bc19d759

Refactor dense FP8 tensor/channel/block utils and add CT FP8 block (#21404)

mgoin committed 81 days ago

Verified fbd6523a

[Structured Output][Refactor] Move `apply_grammar_bitmask()` method from `ModelRunner` to structured output utils (#21999)

shen-shanshan committed 81 days ago

Verified 470484a4

[Misc] Clean up flags in `vllm bench serve` (#25138)

ywang96 committed 81 days ago

Verified 21da7334

[Bugfix][Mamba] - Fix Conv State Kernel FP32 Support (#24883)

Josephasafg committed 81 days ago

Verified 66072b36

Fix `validate-config` pre-commit check (#25157)

hmellor committed 81 days ago

Verified 3ed1ec4a

Fix forward reference warning in documentation (#25150)

hmellor committed 81 days ago

Verified 5a33ae9a

[Docs] add the parallel sampling usage in LLMEngine and AsyncLLM (#24222)

gigit0000 committed 81 days ago

Verified c9ff9e6f

[Docs] Fix pooling-params doc references in openai_compatible_server.md (#24939)

yankay committed 81 days ago

Verified eaffe448

Move `StructuredOutputsConfig` from `config/__init__.py` to `config/structured_outputs.py` (#25153)

hmellor committed 81 days ago

Verified 8ed039d5

[Model] Improve Pooling Model (#25149)

jeejeelee committed 81 days ago

Verified 37970105

Older