vllm-project/vllm

Pull Requests Commits

Merge branch 'main' into enable-stepfun37-testing

jeejeelee committed 3 hours ago

Verified b354aec3

jeejeelee committed 3 hours ago

87adbb0a

[Bug] Fix gemma4 MTP IMA issue when TP>1, `CUDA error: an illegal memory access was encountered` (#43909)

yewentao256 committed 5 hours ago

Verified e1105064

[Governance] Add @BugenZhao as Rust frontend code owner (#44047)

BugenZhao committed 5 hours ago

Verified 50c80d79

[ROCm] Add attention sink support to AITer flash attention backend (#43817)

sphinx07 committed 9 hours ago

Verified 3becc5db

[Bugfix] Fix RMSNorm kernels to multiply in weight's native dtype (#42379)

liulanze committed 13 hours ago

Verified 124fac10

[BugFix][Platform] Fix import vllm.platforms.rocm error on non-CUDA test_gpt_oss.py (#43571)

Liangliang-Ma committed 13 hours ago

Verified e9499996

[ROCm] cmake: support PYTORCH_FOUND_HIP for torch 2.13 native HIP language support (#43881)

nemanjaudovic committed 14 hours ago

Verified c0056b19

[ROCm][CI] Fix failure in the Phi3V pooling test (#44028)

AndreasKaratzas committed 15 hours ago

Verified ef8840ad

[Refactor] Remove dead current_tool_name_sent assignments from tool parsers (#43997)

sfeng33 committed 18 hours ago

Verified 1a096d82

offload prompt_embeds decode in render_prompts_async to avoid blocking (#43792)

gagandhakrey committed 18 hours ago

Verified 1e2ce5d1

[PERF]MiniMax-M2 gate kernel (#38445)

jeejeelee committed 18 hours ago

Verified 559d6710

Revert "[MoE Refactor] Migrate MoeWNA16Method quantization to MK orac… (#44033)

bnellnm committed 20 hours ago

Verified 187457a9

[CI] Fix smoke test step key to bypass block gate (#43974)

khluu committed 20 hours ago

Verified 8fad2665

[CI] Remove duplicate Harmony test coverage (#44023)

sfeng33 committed 21 hours ago

Verified 8c6daf6e

[MoE Refactor] Remove supports_expert_map (#43108)

bnellnm committed 22 hours ago

Verified 7b98f498

[MoE Refactor] Migrate MoeWNA16Method quantization to MK oracle (#42647)

bnellnm committed 22 hours ago

Verified 106aa92f

[Fronten] Clean up stop_token_ids override for Harmony (#44009)

yzong-rh committed 23 hours ago

Verified 46409fd2

[Metrics] Exclude KV transfer tokens from iteration_tokens_total (#43346)

tlrmchlsmth committed 23 hours ago

Verified 38b864d8

[Feature] SSL support for dp supervisor (#43688)

yewentao256 committed 1 day ago

Verified 5dbf1605

Add @khluu to CODEOWNERS (#44019)

khluu committed 1 day ago

Verified acbc2033

[CI] Remove redundant test_chat_with_tool_reasoning.py (#44011)

sfeng33 committed 1 day ago

Verified 6de08e8b

[CI] Make Model Executor test hangs fail fast with a traceback (#43971)

khluu committed 1 day ago

Verified 6aabe221

[Bug] Fix torch device issue for MOE permute (#44005)

yewentao256 committed 1 day ago

Verified 739096a0

[Bugfix] Fix Ray placement group allocation with grouped nodes (#43998)

czhu-cohere committed 1 day ago

Verified 8b9deeec

[Bugfix] Use storage_block_size in KV cache reshape for compressed specs (DeepSeek V4) (#43988)

zixi-qi committed 1 day ago

Verified d07ad069

[EPLB] Make async EPLB default (#43219)

ilmarkov committed 1 day ago

Verified 4aaba00f

[MoE Refactor] WNA16 MoE backend selection into oracle module (#42553)

bnellnm committed 1 day ago

Verified 84b2a8a7

jeejeelee committed 1 day ago

02f61c30

[Bugfix] Disable allreduce_rms_fusion when pipeline_parallel_size > 1 (#43616)

zixi-qi committed 1 day ago

Verified 4ff865c3

Older