vllm-project/vllm

Pull Requests Commits

NickLucche committed 2 days ago

f555f5a1

[Bugfix] Replace assert with ValueError for response_format validation in completions endpoint (#35456)

umut-polat committed 2 days ago

Verified b66a7464

[Bugfix] Use 'sum' reduction instead of 'avg' in Async TP reduce-scatter (#33088)

wangxingran222 committed 2 days ago

Verified 07bdabef

[Model Performance] Add Qwen3MoE tuned MoE configs for H200 (#35457)

chengyinie committed 2 days ago

Verified a572baff

[Bug] correct out dtype of rms_norm_gated native path (#35369)

zufangzhu committed 2 days ago

Verified 516cf266

[Bugfix] disable allreduce_rms_fusion by default when pp size > 1 (#35424)

ZJY0516 committed 2 days ago

Verified 487e5c51

[BugFix] Repo utils debug print patch (#35434)

pi314ever committed 2 days ago

Verified 1a8c7167

[Bug] Fix outdated links in source code (#35314)

yewentao256 committed 2 days ago

Verified 062b7896

use 'max_active_experts' for moe lora input size (#33197)

gnovack committed 2 days ago

Verified a532c838

[Bugfix] Fix Qwen3NextForCausalLM packed_modules_mapping (#35413)

jeejeelee committed 2 days ago

Verified 1e5ad9b7

[Misc] Move `GPUModelRunner.prepare_kernel_block_sizes` to utils (#35400)

NickLucche committed 2 days ago

Verified cabdaa76

[Core]Extract is_last_rank in Ray for tpu to override (#33012)

Chenyaaang committed 2 days ago

Verified 06be5356

[compile] Invalidate cache for cpu flags (#35119)

angelayi committed 2 days ago

Verified c29ee9c3

[Bugfix] Emit reasoning_part events in simple streaming path for Resp… (#35184)

daniel-salib committed 2 days ago

Verified d43048ce

[CI] Actually run tests/kernels/quantization/test_block_fp8.py in CI (#34274)

mgoin committed 2 days ago

Verified 4fec53cf

[Performance] Cublas Bf16 Gate with Fp32 Output (#35121)

roikoren755 committed 2 days ago

Verified 38c498b8

[Update] Use FlashInfer fast_decode_plan directly instead of replication (#34687)

askliar committed 2 days ago

Verified 56a63717

[Bugfix] Fix KV Scale loading for MLA Models (#35430)

pavanimajety committed 2 days ago

Verified 62830211

[ROCm][Quantization] GPT OSS Upstream MoE wmxfp4_afp8 with static scales (#30357)

maleksan85 committed 2 days ago

Verified 01923eec

[Kernel][perf] optimize NCCL symm_mem vs custom_AR selection thresholds (#33839)

pkousha committed 2 days ago

Verified 31fb6f43

[WideEP] Remove pplx all2all backend (#33724)

tlrmchlsmth committed 2 days ago

Verified eb19955c

[Bugfix] Fix MessageQueue connect_ip for cross-node data parallelism (#35429)

luccafong committed 2 days ago

Verified 0f2f24c8

add mixed precision support for modelopt (#35047)

sychen52 committed 2 days ago

Verified d0105b84

Nemotron: use per-layer config in NemotronHMLPDecoderLayer for heterogeneous models (#35396)

danielafrimi committed 2 days ago

Verified 832a780f

[Performance] Extract KV cache update op from flashinfer forward (#35422)

ElizaWszola committed 2 days ago

Verified 98217b09

fix(reasoning): Qwen3ReasoningParser returns truncated output as reasoning (#35230)

stakeswky committed 2 days ago

Verified 967572dd

[Model Runner V2] Prepare attn metadata in ModelState [2/N] (#35383)

WoosukKwon committed 2 days ago

Verified 3d66502e

[Model Runner V2] Add model states [1/N] (#35350)

WoosukKwon committed 2 days ago

Verified c66aa48e

[Model Runner V2] Fix error-handling (#35063)

njhill committed 2 days ago

Verified b6d5a172

[Bugfix] Remove erroneous lower bound on LoRA vocab size constraint (#35354)

LucasWilkinson committed 2 days ago

Verified 5e58bdc7

Older