Commits vllm-project/vllm

[Bugfix] Surface SamplingParams.verify errors as client errors (HTTP 400)

khluu committed 46 days ago

a015ed9c

[CI][Bugfix] Limit MRV2 sampler JIT warmup registration to ROCm (#58465)

khluu committed 46 days ago

Verified 21ee743f

[ROCm][Bugfix] Fix TileLang mHC fused RMSNorm on 64-wide wavefronts (#58419)

djramic committed 46 days ago

Verified 5fcc6e7c

[ROCm][Test] Cover MoRI graph replay and output lifetime (#58093)

AndreasKaratzas committed 46 days ago

Verified 8b660ce9

[Docs] Fix docstring typos (output_dytpe, kwrags, Abbrivations) (#55936)

simpleqt committed 46 days ago

Verified 97b1b121

[Bugfix] Pick a KV block size supported by every attention backend (#49845)

divyvasal committed 46 days ago

Verified e26547ad

[CI] Select one GPU for the H200 initialized snapshot E2E step (#58351)

Thangnguyenvn98 committed 46 days ago

Verified 34b06908

[Perf] Use breakable CUDA graphs (no torch.compile) by default under VLLM_BATCH_INVARIANT so the tuned matmul configs see the runtime M (#57586)

LioEinaudi committed 46 days ago

Verified 711fc55c

Doc: add DiffusionGemma to supported models (#46466)

BPbruce committed 46 days ago

Verified f9ad9dd6

[Bugfix][ROCm] Fix startup OOM in AITER MLA FP8 prefill workspace sizing (#57923)

simondanielsson committed 46 days ago

Verified b5ea0c71

[Bugfix][NIXL] Restore successful push completion reporting (#58188)

Dao007forever committed 46 days ago

Verified c4f6ce48

[CI][ROCm] Add an MI355 Kimi-K3 unit test group (#58012)

okorzh-amd committed 46 days ago

Verified c4cfd700

[ROCm][CI] Validate Mooncake and NIXL prefill/decode accuracy (#58095)

AndreasKaratzas committed 46 days ago

Verified 8033bd08

[ROCm][Perf] Extend QK-norm/RoPE/KV-cache fusion to MRoPE (#50212)

vorapolsiloai committed 46 days ago

Verified b715699d

[Quark] Remove quark-specific silent online quantization (#51800)

fxmarty-amd committed 46 days ago

Verified 0549e8d0

[Quantization][Bugfix] Bump humming-kernels to 0.1.16 (#58054)

jinzhen-lin committed 46 days ago

Verified f69cc7c1

[Quantization] Enable humming wNaM asymmetric quant (zero_point) with compressed-tensors (#46528)

HDCharles committed 46 days ago

Verified 9f07d023

[CPU] Adds support for fp32 attention sinks (#56252)

Ankit-Jaiswal-AMD committed 46 days ago

Verified 88afb9dc

[Perf][MoE] Skip top-k slots routed to non-local experts in TritonExp… (#58051)

ShuoleiWang committed 46 days ago

Verified 9ed81581

[Feature][Frontend] Add DeepSeek-V4 FIM completion rendering (#44229)

QwertyJack committed 46 days ago

Verified d95a896e

[DSpark] Support pipeline-parallel targets in aggregated serving (#56956)

lucifer1004 committed 46 days ago

Verified 157bcb7c

[Bugfix][Tool Parser] Migrate Granite to the streaming Parser Engine (#49648)

nikhilkulkarni1755 committed 46 days ago

Verified 15ed1262

[Bugfix] Disable prefix caching for encoder-only before model config hooks (#58287)

gty111 committed 46 days ago

Verified c9611215

[ROCm][Test] Check GDN prefill numerics and output ownership (#58091)

AndreasKaratzas committed 46 days ago

Verified a22c637e

[XPU] upgrade to PyTorch 2.14 (#56013)

yma11 committed 46 days ago

Verified e34489a5

[Bugfix][SM120][MLA] Support NoPE sparse MLA (GLM-5.3-Flash) on the FlashInfer SM120 backend (#55277)

lucifer1004 committed 46 days ago

Verified c843f0ca

[MRV2] Miscellaneous code cleanup (#57980)

njhill committed 46 days ago

Verified afec265b

[ROCm][CI] Include Python tooling in ROCm CI artifacts (#58271)

AndreasKaratzas committed 46 days ago

Verified 36bc6948

[ROCm][Bugfix] Keep zero MiniMax MXFP8 activation blocks finite (#58089)

AndreasKaratzas committed 46 days ago

Verified 56b3acb5

[Perf][ROCm][Attention] Narrow the Triton prefill-attention KV tile on RDNA3/RDNA4 (#58225)

lijipeng787 committed 46 days ago

Verified 1904ed94