vllm-project/vllm

Pull Requests Commits

Bump actions/setup-python from 5.4.0 to 5.6.0

dependabot[bot] committed 27 days ago

Verified 3ed73a0f

[Minor][Models] Pass partial_rotary_factor parameter to rope (#17266)

Eviannn committed 27 days ago

Verified 7fcc4223

[Misc] Validate `stop_token_ids` contents (#17268)

njhill committed 27 days ago

Verified 8262a3e2

[Doc] small fix (#17277)

reidliu41 committed 27 days ago

Verified f211331c

[Doc] Fix wrong github link in LMCache examples (#17274)

KuntaiDu committed 27 days ago

Verified 9053d0b1

[Bugfix] Fix Mistral3 spatial merge error (#17270)

mgoin committed 27 days ago

Verified cb3f2d8d

[Bugfix] Fix cutlass dispatch for fp8/int8 to properly invoke M<=16 c… (#16751)

Ther-LF committed 27 days ago

Verified c12df53b

[Bugfix] Fix missing ARG in Dockerfile for arm64 platforms (#17261)

lkm-schulz committed 27 days ago

Verified d1aeea75

[BugFix] Fix vllm_flash_attn install issues (#17267)

LucasWilkinson committed 27 days ago

Verified d8bccde6

[V1][Spec Decode] Make eagle compatible with prefix caching. (#17137)

LiuXiaoxuanPKU committed 27 days ago

Verified 20e489ea

[Metrics] Fix minor inconsistencies in bucket progression (#17262)

DarkLight1337 committed 27 days ago

Verified 4213475e

[doc] Add feature status legend (#17257)

reidliu41 committed 27 days ago

Verified d92879ba

[Feature] support sequence parallelism using compilation pass (#16155)

cascade812 committed 27 days ago

Verified 690fe019

[NVIDIA] Support Cutlass MLA for Blackwell GPUs (#16032)

kaixih committed 27 days ago

Verified ed7a29d9

[Bugfix] Fix Lora Name Parsing (#17196)

alex-jw-brooks committed 28 days ago

Verified 756848e7

[Misc] Change buckets of histogram_iteration_tokens to [1, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8096] to represent number of tokens (#17033)

sfc-gh-zhwang committed 28 days ago

Verified 18445edd

[MISC] Use string annotation types for class definitions (#17244)

jianzs committed 28 days ago

Verified 30215ca6

[Bugfix] Get a specific type of layer from forward context (#17222)

heheda12345 committed 28 days ago

Verified 838cedad

[Bugfix] Fix QWen2 VL multimodal mapping (#17240)

jeejeelee committed 28 days ago

Verified 4283a28c

[Misc] Make cached tokenizer pickle-compatible (#17048)

DarkLight1337 committed 28 days ago

Verified 93a126fb

[Kernel][Triton][FP8] Adding fp8 and variable length sequence support to Triton FAv2 kernel (#12591)

rasmith committed 28 days ago

Verified 8e4b351a

Update test_flash_attn.py (#17102)

ShuaibinLi committed 28 days ago

Verified 9869453c

[CI/Build] remove -t for run-lm-eval-gsm-hf-baseline.sh (#16271)

reidliu41 committed 28 days ago

Verified 3642c59a

[Minor] Fix lint error in main branch (#17233)

WoosukKwon committed 28 days ago

Verified 43eea295

[Bugfix] Fix Qwen2.5-Omni M-RoPE position ids generation (#16878)

imkero committed 28 days ago

Verified de7eb10c

[MISC] rename interval to max_recent_requests (#14285)

andyxning committed 28 days ago

Verified fd11a325

Disable the torch.compile cache checks when VLLM_DISABLE_COMPILE_CACHE=1 (#16573)

houseroad committed 28 days ago

Verified 4d17e203

[Bugfix] fix error due to an uninitialized tokenizer when using `skip_tokenizer_init` with `num_scheduler_steps` (#9276)

junstar92 committed 28 days ago

Verified 10fd1d73

[Docs] Update structured output doc for V1 (#17135)

russellb committed 28 days ago

Verified 52b4f4a8

[Chore] added stubs for `vllm_flash_attn` during development mode (#17228)

aarnphm committed 28 days ago

Verified e782e0a1

Older