Commits vllm-project/vllm

[CI] Allow empty final audio streaming chunk

khluu committed 55 days ago

6c11f7f8

[CI][AMD] Honor single-node Docker workload timeout (#52547)

tjtanaa committed 55 days ago

Verified d569e71b

[Rust Frontend][RL]: report engine world size over gRPC (#53204)

biswapanda committed 55 days ago

Verified 9dba2c9b

[ROCm][CI] Restore attention coverage after KV-cache layout refactor (#53351)

AndreasKaratzas committed 55 days ago

Verified 30b34171

[Deprecation] Remove dead use_prefill_decode_attention flag (#52557)

brianosaurus committed 55 days ago

Verified bbe8b23e

[Model] Add Qwen3-Omni DSpark support (#52560)

Zhou248 committed 55 days ago

Verified 2f55ef25

[Pooling] Fix Rust pooling endpoint for benchmark (#53352)

taneem-ibrahim committed 55 days ago

Verified a014e35f

[Mypy Fix] Mypy fix for "vllm/model_executor/models/[eE][fF]" (#53381)

yewentao256 committed 55 days ago

Verified 811d12ee

[Core][RL] Support sparse checkpoint updates through native weight loaders (#50723)

ShuoleiWang committed 55 days ago

Verified 236f78cc

[Perf] Support internal prefill checkpoints for Mamba prefix caching, 9%~25% TTFT improvement (#52789)

yewentao256 committed 55 days ago

Verified 9eb9d9d3

[MM] Remove renderer_applies_updates flag (#53385)

DarkLight1337 committed 55 days ago

Verified 020829f6

[ROCm] Ship rocprofiler-sdk 1.3.2 in Dockerfile.rocm_base to fix torch.profiler traces (#53182)

Rohan138 committed 55 days ago

Verified 610bfc58

[Bugfix] Fix speculative decoding for short_conv (LFM2) models (#50272)

zwischenraum committed 56 days ago

Verified da329cc3

[Pooling] Report input throughput for batched requests (#53213)

taneem-ibrahim committed 56 days ago

Verified 704f12aa

[CI][ARM64] Restore known-good CUDA 13 builder image (#53374)

khluu committed 56 days ago

Verified 10704541

[MM] Address comments on #53275 (#53364)

DarkLight1337 committed 56 days ago

Verified 040700aa

[MM] Simplify _apply_hf_processor_main (#53275)

DarkLight1337 committed 56 days ago

Verified 0b19ebca

[Refactor][Model Runner V2][Multimodal] Move the encoder-only path out of the shared runner (#53176)

gty111 committed 56 days ago

Verified 7ca49fbe

[Bugfix][Kimi K3] Enable deferred MoE finalization before weight loading (#53327)

zyongye committed 56 days ago

Verified e9d1398d

[Bugfix][R3] Unwrap UniformTypeKVCacheSpecs when selecting the routed-experts KV group (#53240)

HollowMan6 committed 56 days ago

Verified b2db227a

[Bugfix][ROCM] Fix the MXFP8 block scale exponent (#53110)

stefankoncarevic committed 56 days ago

Verified 7f4a1b7e

[6/N][KV-Cache Layout Refactor] Standardize KV cache layout (#51718)

LucasWilkinson committed 56 days ago

Verified 8bdc70ec

[Bugfix][Spec Decode] Use group geometry for FlashAttention metadata (#53002)

mgoin committed 56 days ago

Verified 9ff7041b

[Test] Add focused hybrid MTP prefix-cache regressions (#53189)

mgoin committed 56 days ago

Verified a34f2ab7

Forward Anthropic vllm_xargs to sampling params (#53308)

vMaroon committed 56 days ago

Verified a556f3fc

Fix Cohere ChatV2 citation and tool handling issues (#52175)

andrewbcohere committed 56 days ago

Verified f94dcde5

[CI/Build][ROCm] Run the TileLang HIP symbol checks in their own interpreter (#53117)

stefankoncarevic committed 56 days ago

Verified f37d5868

Revert "Remove native Hunyuan V1 and VL implementations" (#53296)

khluu committed 56 days ago

Verified e3f60265

Revert "[ROCm][Perf] Kimi-K3 Fused kernels for KDA prefill" (#53294)

khluu committed 56 days ago

Verified 592e06f2

[Bugfix][KV Connector][NIXL] Support PCP producers (#52779)

LucasWilkinson committed 56 days ago

Verified 0a3a4f7f