vllm-project/vllm

Pull Requests Commits

[ROCm][Bugfix] Fix `fa_version` argument error in `flash_attn_maxseqlen_wrapper` for ROCm without aiter (#30909)

AndreasKaratzas committed 2 days ago

Verified 8da6ae49

[BugFix] Partial revert of #29558 (DeepEP HT + PIECEWISE CG support) (#30910)

LucasWilkinson committed 2 days ago

Verified 30bb19a7

[Bugfix] Fix Unicode issues in GLM-4 tool calling (#30920)

chaunceyjiang committed 2 days ago

Verified aa7e8360

[ROCm][Bugfix] fix(structured_output): Skip guidance backend for schemas with patternProperties (#30730)

AndreasKaratzas committed 2 days ago

Verified be2ad5f9

[Platform] Let EPD work with non-cuda platform (#30225)

wangxiyuan committed 2 days ago

Verified a85724bd

[Fix][FlexAttention] return max logical block index to handle reused blocks (#30915)

ivanium committed 2 days ago

Verified 11a89cf9

[CPU] Refactor CPU fused MOE (#30531)

bigPYJ1151 committed 2 days ago

Verified e3ab93c8

fix: add warmup for audio preprocessing (#30706)

TheCodeWrangler committed 2 days ago

Verified fc2ae6d6

[KV connector][LMCache] Only record the cuda event when there are request to store/load (#30814)

ApostaC committed 2 days ago

Verified ec965569

[AMD][CI] fix lm eval ci arg (#30911)

divakar-amd committed 2 days ago

Verified 82dc338a

[PERF] Qwen3-next. Add fp8 cutlass MoE tuned configs. `chmod -x *MI308X.json` (#29553)

vadiklyutiy committed 2 days ago

Verified 717ac33d

[Doc][CPU] Update CPU doc (#30765)

bigPYJ1151 committed 2 days ago

Verified cfb7e555

[refactor] Add prefix support to embed_tokens in DeepSeek MTP (#30788)

zzhx1 committed 2 days ago

Verified b166ef20

[compile] Fix CI for test_gpt2_cache_hit (#30902)

zhxchen17 committed 2 days ago

Verified 5f2f3fba

[UX] Reduce DeepGEMM warmup log output to single progress bar (#30903)

MatthewBonanni committed 2 days ago

Verified 4a8412f7

[Quantization] Support Quark int4-fp8 w4a8 for MoE (#30071)

BowenBao committed 2 days ago

Verified 0c738b58

fused_moe_lora PDL improvements (#30716)

gnovack committed 2 days ago

Verified 5a3adf58

[Chore] Remove v0 dead code for Qwen2.5-omni (#30883)

Isotr0py committed 2 days ago

Verified 6fe58876

[NIXL] Support P tensor-parallel-size > D tensor-parallel-size (#27274)

NickLucche committed 2 days ago

Verified bc3700e0

[ROCm][CI] Reduce Flakiness For test_async_scheduling Using ROCM_ATTN With FP32 (#30811)

micah-wil committed 2 days ago

Verified fd8afdf3

[Metrics] Model FLOPs Utilization estimation (#30738)

SungMinCho committed 2 days ago

Verified a0b782f9

Update vllm/model_executor/models/qwen2_moe.py

yewentao256 committed 2 days ago

Verified 710f6e44

[CI][Feature] Adds auto-rebase PR rule (#30875)

rafvasq committed 2 days ago

Verified ed2897f3

fix torch inductor issue

yewentao256 committed 2 days ago

26e09a1c

[v1] Add PrefixLM support to TritonAttention backend (#30386)

Isotr0py committed 2 days ago

Verified 74a1ac38

feat(api): Eager chat template warmup to eliminate first-request latency (#30700)

TheCodeWrangler committed 2 days ago

Verified 05a83dc6

[BugFix] Workspace allocation during profile run : DeepEPHighThroughput + DeepGEMM (#30899)

varun-sundar-rabindranath committed 2 days ago

Verified e3fc374a

2.9.1 PyTorch release update (#28495)

atalman committed 3 days ago

Verified e06d0bf0

[docs]: add ecosystem projects sr in docs/governance (#30844)

Xunzhuo committed 3 days ago

Verified e3a0f21e

[Attention] Update tests to remove deprecated env vars (#30563)

MatthewBonanni committed 3 days ago

Verified 7eb6cb6c

Newer Older