vllm-project/vllm

Pull Requests Commits

Merge branch 'main' into codex/rocm-artifact-tensorizer

AndreasKaratzas committed 15 days ago

Verified 78d2334a

[Model Runner V2] Migration to support quantized model by default [5/N] (#44446)

yewentao256 committed 15 days ago

Verified b53b1c7f

[Bugfix][Quantization] Don't reject fp8_e5m2 KV cache for non-fp8 quantized checkpoints (#45040)

Sunt-ing committed 16 days ago

Verified 79ca54d2

[Bugfix] [Parser] Fix Qwen3 latent bug in partial params dropping values containing `<` (#46047)

bbrowning committed 16 days ago

Verified 09f3cd5c

[KV Connector][Offloading] Disable parallel-agnostic fs-tier cache on V2 model runner (#46044)

Etelis committed 16 days ago

Verified ea6078fe

[Tests] Add Qwen3 streaming parser delta boundary cases (#45708)

Palaiologos1453 committed 16 days ago

Verified a0df04e4

[ROCm][Spec Decode] Fix probabilistic draft probs test attention backend (#45706)

stefankoncarevic committed 16 days ago

Verified e2352c29

[CI]Enable mxfp4 lora test for ROCm platform (#43802)

qli88 committed 16 days ago

Verified 25faa1f4

[Bugfix][Kernel] Check output alignment in vectorize_with_alignment (fixes misaligned-address crash for non-multiple-of-8 head sizes) (#45466)

HumphreySun98 committed 16 days ago

Verified 4583630b

[ROCm][CI] move lora%N test to mi300 and gate (#45970)

divakar-amd committed 16 days ago

Verified 21da47da

[Frontend] Add Streaming Parser Engine and new GLM4.7/GLM5.1/GLM5.2 Parser (#45915)

chaunceyjiang committed 16 days ago

Verified 6c379b9e

[Bugfix][ROCm] Fix rocm_aiter_per_tensor_quant custom op aliasing (#45747)

Rohan138 committed 16 days ago

Verified 50994746

[Bugfix] Restore is_sym guard for zp in GPTQ/CT MoE to fix symmetric quant regression (#45656)

yuwenzho committed 16 days ago

Verified 058cc0a8

[Bugfix][Tool Parser] Handle non-finite numbers in coerce_to_schema_type (#43984)

ashishpatel26 committed 16 days ago

Verified 837db760

Temporarily remove @markmc from CODEOWNERS (#46053)

markmc committed 16 days ago

Verified bf2a3930

[Model] Remove BambaForCausalLM (#45990)

xianbaoqian committed 16 days ago

Verified d682968a

Fix _riscv_supports_rvv_vlen128() to detect RVV on hardware without zvl flags (#43179)

lyd1992 committed 16 days ago

Verified 021cdf72

[Rust Frontend]: Add `/get_world_size` route with static parallel size (#44801)

coder3101 committed 16 days ago

Verified 4cb5e746

[Kernel] Add PDL support for DeepGEMM kernel (#46006)

jeejeelee committed 16 days ago

Verified 22cc8911

[ROCm][DSv4] Functional fixes for DeepSeek V4 on MI300X/MI325X (#45681)

tuukkjs committed 16 days ago

Verified afdcbd5d

fix(quantization): Fix AWQ dequantize on Intel XPU and refactor AutoAWQ config (#42727)

Alex-ai-future committed 16 days ago

Verified 8d4f5496

[CPU] Skip Triton kernel monkey-patches when Triton-CPU is available (#44991)

jmamou committed 16 days ago

Verified 351c72d6

[Rust Frontend] Return model metadata fields in /v1/models (#45950)

tahsintunan committed 16 days ago

Verified 7299e650

Fix Stale Encoder Cache After Weight Update (#45093)

littlecircle0730 committed 16 days ago

Verified 08985351

[Mooncake] Skip KV lookup for non-reachable SWA blocks (#45444)

wzhao18 committed 16 days ago

Verified 5fd3b276

fix(anthropic): preserve inline system message position for prefix caching (#44602)

felix0080 committed 16 days ago

Verified 1e9f04da

[Bugfix][Frontend] Fix Anthropic count_tokens decorator order driving server load negative (#44725)

Sunt-ing committed 16 days ago

Verified 70221414

[XPU] Update nixl to v0.10.1 in Dockerfile (#40287)

zhenwei-intel committed 16 days ago

Verified a3315893

Revert "[Kernel] Add PDL support for DeepGEMM kernel" (#45999)

micah-wil committed 16 days ago

Verified e9451692

[Test][KV Connector] Add request_finished fence population tests for offloading scheduler (#45679)

Alex-ai-future committed 16 days ago

Verified 554352a3

Older