[BugFix] Fix bad words with speculative decoding (#31908)

Nick Hill committed 10 months ago

[EPLB] Optimize EPLB with numpy (#29499)

Ilya Markov committed 10 months ago

[Kernel] Support bias type in grouped_topk kernel (#31781)

Xin Yang committed 10 months ago

[refactor] refactor memory constants usage (#31865)

Ning Xie committed 10 months ago

[Perf] Fuse stride preparation for NVFP4 cutlass_moe (#31837)

Michael Goin committed 10 months ago

[Doc] Fix: Correct vLLM announcing blog post link in docs (#31868)

Festus Ayobami Owumi committed 10 months ago

Enable quantized attention in NemotronH models (#31898)

roikoren755 committed 10 months ago

UX: add vLLM env info in '/server_info' (#31899)

Jee Jee Li committed 10 months ago

[KVConnector]: Enable Cross-layers KV cache layout for MultiConnector (#30761)

Kfir Toledo committed 10 months ago

[Bugfix]: prevent leaking tokens in crash log (#30751)

Marko Rosenmueller committed 10 months ago

[Refactor] Clean up pooler modules (#31897)

Cyrus Leung committed 10 months ago

[Perf][Kernels] Enable FlashInfer DeepGEMM swapAB on SM90 (for W8A8 Linear Op) (#29213)

Kate Cheng committed 10 months ago

[OpenAI] Extend VLLMValidationError to additional validation parameters (#31870)

R3hankhan committed 10 months ago

[Chore] Migrate V0 attention utils (#31891)

Cyrus Leung committed 10 months ago

[Refactor] GLM-ASR Modeling (#31779)

Jared Wen committed 10 months ago

[ROCm][AITER] fix wrong argument passed to AITER `flash_attn_varlen_func` (#31880)

vllmellm committed 10 months ago

[Bugfix][MTP] Fix GLM4 MoE fp8 loading with MTP on (#31757)

Andy Liu committed 10 months ago

[Misc] Improve error messages for unsupported types and parameters (#30593)

BlankR committed 10 months ago

[Model] Cleanup: Remove redundant manual definition of `make_empty_intermediate_tensors` in GLM-4-MoE (#31869)

maang committed 10 months ago

[XPU]fallback to TRITON_ATTN on xpu when use float32 dtype (#31762)

sihao_li committed 10 months ago