vllm-project/vllm

Pull Requests Commits

[Bugfix] Update multimodel models mapping to fit new checkpoint after Transformers v4.52 (#19151)

Isotr0py committed 200 days ago

Verified ca94d7fa

[Mis] remove duplicate engine status checks (#19647)

googs1025 committed 200 days ago

Verified 5a1c2e15

[V1][Kernel] Flashinfer HND KV cache layout (#19280)

NickLucche committed 200 days ago

Verified 4c8f64fa

[doc] split "Other AI Accelerators" tabs (#19708)

davidxia committed 200 days ago

Verified 93aee29f

[doc][mkdocs] Add edit button to documentation (#19637)

reidliu41 committed 200 days ago

Verified 154d063b

[Kernel] Add Split-KV Support to Unified Triton Attention Kernel (#19152)

jvlunteren committed 200 days ago

Verified ccd7c050

Add a doc on how to update PyTorch version (#19705)

huydhn committed 200 days ago

Verified c48c6c40

[Doc] Add missing llava family multi-image examples (#19698)

Isotr0py committed 200 days ago

Verified aed84686

[Core] add remove_seq_from_computed_blocks_tracker to BlockSpaceManager (#19686)

quanliu1991 committed 200 days ago

Verified 5c76b9cd

Fixes IMA for TP w/ flex-attention (#19712)

drisspg committed 200 days ago

Verified ddfed314

[DOC] fix doc typos (#19600)

diliu0349 committed 200 days ago

Verified 5b3ad5ec

[Frontend] add chunking audio for > 30s audio (#19597)

nguyenhoangthuan99 committed 200 days ago

Verified ede5c4eb

[Wheel Size] Only build FA2 8.0+PTX (#19336)

LucasWilkinson committed 200 days ago

Verified 07334959

[doc] add project flag to gcloud TPU command (#19664)

davidxia committed 200 days ago

Verified 119f6839

[Fix] Fall back to Gloo when NCCL backend is unavailable (#19641)

conroy-cheers committed 200 days ago

Verified 0860087a

[Quantization] Remove FP4 emulation; Fall-back to marlin for device < 100 (#19563)

dsikka committed 200 days ago

Verified 6bc7b573

[V1] Change return type on get_multimodal_embeddings() (#19446)

russellb committed 201 days ago

Verified 90f9c2eb

[Model] Add support for MiniMaxM1ForCausalLM (shares architecture with MiniMaxText01ForCausalLM) (#19677)

qscqesze committed 201 days ago

Verified 387bdf0a

[Kernels] Use empty for modular MoE workspaces (#19667)

bnellnm committed 201 days ago

Verified 5e5baa91

[Bugfix] fix missing 'finish_reason': null in streaming chat (#19662)

chaunceyjiang committed 201 days ago

Verified 836d4ce1

[MISC] bump huggingface_hub pkg to 0.33.0 (#19547)

andyxning committed 201 days ago

Verified c3fec47b

[Bugfix] Fix TP inference for Flex attention backend (#19657)

Isotr0py committed 201 days ago

Verified 1173804d

[Feature]:Allow for Granite MoE Hybrid models with _only_ shared experts. (#19652)

shawntan committed 201 days ago

Verified 4d542402

[DOC] Add reasoning capability to vLLM streamlit code (#19557)

Navanit-git committed 201 days ago

Verified 3e750697

[BugFix] Don't catch BaseException when dumping execute_model errors (#19626)

njhill committed 201 days ago

Verified ee35e96a

[Kernel] GGUF MMVQ kernel for multiple input vectors (#18754)

SzymonOzog committed 201 days ago

Verified dec66d25

[Docs] Move multiproc doc to v1 dir (#19651)

russellb committed 201 days ago

Verified 8d120701

[CI] Add mteb testing for rerank models (#19344)

noooop committed 201 days ago

Verified f40f763f

[MISC] typo fix (#19672)

andyxning committed 201 days ago

Verified 26bc46ef

[TPU] support attention head dim smaller than 128 (#19620)

yaochengji committed 201 days ago

Verified a77aea59

Newer Older