vllm-project/vllm

Pull Requests Commits

[TPU][V1] Add support for top-logprobs (#17072)

NickLucche committed 3 days ago

Verified 5941e0b7

[TPU] Enable gemma3-27b with TP>1 on multi-chips. (#17335)

vanbasten23 committed 3 days ago

Verified 97659408

[BugFix] Increase timeout for startup failure test (#17642)

njhill committed 3 days ago

Verified 5ea5c514

[Benchmarks] Remove invalid option under V1 engine (#17651)

russellb committed 3 days ago

Verified d3efde81

Use git-path commit in hook (#17616)

thomasjpfan committed 3 days ago

Verified aea302be

[Doc] Fix broken cuda installation doc rendering (#17654)

Isotr0py committed 3 days ago

Verified cc05b90d

[Kernel] some optimizations for dense marlin and moe marlin (#16850)

jinzhen-lin committed 3 days ago

Verified 1d0c9d6b

[Build/CI] Upgrade CUTLASS to 3.9.2 (#17641)

tlrmchlsmth committed 4 days ago

Verified f62cad64

[Bugfix] fix KeyError on top logprobs are special tokens (#17637)

chaunceyjiang committed 4 days ago

Verified 5394ad73

[Bugfix][Easy] Fix whitespace in shm_broadcast.py logging (#17635)

tlrmchlsmth committed 4 days ago

Verified 68e1ee00

[Bugfix] Prioritize dtype in root config before checking text config (#17629)

DarkLight1337 committed 4 days ago

Verified 2858830c

Add full API docs and improve the UX of navigating them (#17485)

hmellor committed 5 days ago

Verified d6484ef3

[Misc] V0 fallback for `--enable-prompt-embeds` (#17615)

DarkLight1337 committed 5 days ago

Verified 46fae69c

[Bugfix] Fix broken Qwen2.5-omni tests (#17613)

Isotr0py committed 5 days ago

Verified f66f1e0f

[Core] Gate `prompt_embeds` behind a feature flag (#17607)

DarkLight1337 committed 5 days ago

Verified 887d7af8

[Bugfix][ROCm] Using device_type because on ROCm the API is still torch.cuda (#17601)

gshtras committed 6 days ago

Verified a9284245

[Build/CI] Upgrade CUTLASS to 3.9.1 (#17602)

tlrmchlsmth committed 6 days ago

Verified c8386fa6

[Frontend][TPU] Add TPU default max-num-batched-tokens based on device name (#17508)

Chenyaaang committed 6 days ago

Verified 87baebeb

[Quantizaton] [AMD] Add support for running DeepSeek int8 w8a8 MoE on ROCm (#17558)

rasmith committed 6 days ago

Verified e3d0a1d1

Update test requirements to CUDA 12.8 (#17576)

22quinn committed 6 days ago

Verified d47b605e

[Neuron][Build] Require setuptools >= 77.0.3 for PEP 639 (#17603)

liangfu committed 6 days ago

Verified 22c6f639

[release] Add command to clean up Docker containers/images in TPU release machine (#17606)

khluu committed 6 days ago

Verified 3ec97e2c

fix typo in logging (#17605)

ehartford committed 6 days ago

Verified 9b103a1d

[easy] Print number of needed GPUs in skip message (#17594)

zou3519 committed 6 days ago

Verified b90b0852

[Hardware][AMD] Improve OAM device ID + llama4 Maverick MOE tuning (#16263)

xw285cornell committed 6 days ago

Verified 9352cdb5

Add NVIDIA TensorRT Model Optimizer in vLLM documentation (#17561)

Edwardf0t1 committed 6 days ago

Verified 182f40ea

permute/unpermute kernel for moe optimization (#14568)

CalebDu committed 6 days ago

Verified 3e887d2e

[BugFix][Attention] Fix sliding window attention in V1 giving incorrect results (#17574)

LucasWilkinson committed 6 days ago

Verified 0f87d8f7

[Bugfix] fix tmp_out and exp_sums dimensions (#17438)

hliuca committed 6 days ago

Verified 4c33d673

[Misc] Clean up input processing (#17582)

DarkLight1337 committed 6 days ago

Verified cb234955

Newer Older