vllm-project/vllm

Pull Requests Commits

Add tasks/version/logging_utils to models_basic GPU deps

khluu committed 6 days ago

74690b24

Add sequence.py and config/ to models_basic deps

khluu committed 6 days ago

8ac682b5

[CI] Narrow models_basic.yaml source dependencies

khluu committed 6 days ago

19662612

[ROCm] Cap Triton paged attention block size to fix ROCm shared memory OOM (#38502)

AndreasKaratzas committed 6 days ago

Verified 0a309b5e

[CI] Trigger LoRA test when changing MoE code. (#42196)

jeejeelee committed 6 days ago

Verified 84f7a553

[Docs] Fix broken local links (#42160)

chfeng-cs committed 6 days ago

Verified a2c9d548

Add @zyongye to CODEOWNERS (#42200)

zyongye committed 6 days ago

Verified 301305c0

Fix mypy failure on main (#42197)

mmangkad committed 6 days ago

Verified efd0e778

[Frontend][Bugfix] Abort ASR engine requests on cancellation (#41266)

abdulrahman-cohere committed 6 days ago

Verified a5d0a5af

[ROCm][CI] Fix NIXL spec-decode acceptance startup and diagnostics (#41313)

AndreasKaratzas committed 6 days ago

Verified f2840120

[Bugfix][KV Transfer][NIXL] Notify P node on pre-admission rejection to free stranded KV blocks (#41269)

Dao007forever committed 6 days ago

Verified 3f5bd482

[ROCm][CI] Stabilize ROCm shutdown and distributed compile CI (#41573)

AndreasKaratzas committed 6 days ago

Verified fb1ac806

[Bugfix] Fix DeepSeek v4 topk numerical issue for unaligned max-model-len (#42169)

wzhao18 committed 6 days ago

Verified 986edc85

docs: clarify Gemma 4 assistant speculative decoding (#42180)

AbhiOnGithub committed 6 days ago

Verified 27d3bac2

Use CU_MEMCPY_SRC_ACCESS_ORDER_ANY for batch KV cache swaps (#39306)

Etelis committed 6 days ago

Verified 00b0618a

Handle optional bool-or-string CLI args in get_kwargs (#40951)

cvan20191 committed 6 days ago

Verified 0d382ecd

[CI/Build] Use modelscope's international site for regression test (#42176)

Isotr0py committed 6 days ago

Verified 1029e5ef

[Bugfix] Fix SP pass for multimodal models and PP+SP residual handling (#33322)

wangxingran222 committed 6 days ago

Verified 0b272a6e

Fix: Nemotron 3 rescue whitespace-only final_content, not just None (#41846)

Naveassaf committed 6 days ago

Verified dcb3135a

Add NVFP4 all-gather GEMM fusion for AsyncTP (#41882)

baonudesifeizhai committed 6 days ago

Verified bc5fdc1e

[Bugfix] Skip routed-experts hot path when disabled (#42148)

aoshen02 committed 6 days ago

Verified 006af4b9

[KV Connector] Remove compat support for pre-v0.12.0 constructor signatures without `KVCacheConfig` (#39832)

yewentao256 committed 6 days ago

Verified ea0e501b

[Refactor] Nixl util using lazy init (#41392)

yewentao256 committed 6 days ago

Verified f80aa53c

[Quantization] Add ModelOpt NVFP4 W4A16 (4-bit weights, fp16/bf16 activations) support (#41769)

juhi10071998 committed 6 days ago

Verified 7a2b5969

[SpecDecoding] extend mtp support for mimo 2.5 (#41905)

ZJY0516 committed 6 days ago

Verified 2ee8c2a5

[Model] use AutoWeightsLoader for DeepSeekV2 (#41706)

SoluMilken committed 6 days ago

Verified cd74911d

[BugFix] Fix Gemma4 'layers.0.moe.experts.0.down_proj_packed' KeyError issue (#40708)

SoluMilken committed 6 days ago

Verified 25abddc1

[Bugfix][PD] Fix DSv4 Disaggregated (#41957)

NickLucche committed 7 days ago

Verified 171d59ae

[Bugfix] Remove nested torch.compile in GDN rearrange_mixed_qkv causing CUDA graph capture failure (#42070)

tdoublep committed 7 days ago

Verified 3dda9aeb

[Bugfix] Fix GDN KKT precision loss on Hopper GPUs by aligning tl.dot operand layout with WGMMA (#42076)

Kermit-C committed 7 days ago

Verified adb6d965

Older