vllm-project/vllm

Pull Requests Commits

Revert "[Perf] Batch KV cache swap copies via cuMemcpyBatchAsync (#38460)"

khluu committed 58 days ago

3da0bdec

[Bugfix] Re-enable Renormalize routing for TRT-LLM MoE experts (#38859)

yzong-rh committed 58 days ago

Verified a5a623d9

[vLLM IR] add `import_ir_kernels()` to support OOT platforms (#38807)

wxsIcey committed 58 days ago

Verified f8c3af2d

Fix invalid logprobs with MTP enabled and sync scheduling (#38711)

danisereb committed 58 days ago

Verified 50cd5674

[Frontend] new online quantization frontend (#38138)

vkuzo committed 59 days ago

Verified 7b1a7423

[KVConnector] Skip `register_kv_caches` on profiling (#38558)

NickLucche committed 59 days ago

Verified 97f92c6b

[Bugfix] Fix AWQ models batch invariance issues (#38670)

YM2132 committed 59 days ago

Verified 46f02e00

[XPU] bump up xpu-kernel v0.1.5, transpose moe weights (#38342)

mayuyuace committed 59 days ago

Verified 6b487224

[Kernel] Add swapAB support for SM120 CUTLASS blockwise FP8 GEMM (#38325)

Nekofish-L committed 59 days ago

Verified 580090db

[GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill (#38361)

arpera committed 59 days ago

Verified cb10b7e8

[Intel][Triton] Support `round_int8` for Intel backend (#38825)

mieshkiwrk committed 59 days ago

Verified bf8b022e

[XPU][CI] Skip test_topp_only and test_topk_and_topp cases on Intel GPU in CI (#38904)

zxd1997066 committed 59 days ago

Verified 40ee64c0

[ROCm] Fix aiter persistent mode mla with q/o nhead<16 for kimi-k2.5 tp8 (#38615)

wufann committed 59 days ago

Verified 1b117cb0

[CPU] Replace OMP initialization (#36487)

kot-begemot-uk committed 59 days ago

Verified abebd932

[Frontend] feat: add streaming support for token generation endpoint (#37171)

hhk7734 committed 59 days ago

Verified 25f2b553

[XPU][CI] Skip test_topk_only cases on Intel GPU in CI (#38899)

zxd1997066 committed 59 days ago

Verified cb4ff07f

[ROCm][CI/Build] Fix the pytest hook to properly print out the summary (#38585)

gshtras committed 59 days ago

Verified a7d79fa1

Fix Nano Nemotron VL regressions (#38655)

netanel-haber committed 59 days ago

Verified fa9e6802

[Misc] Clean up Gemma4 implementation (#38872)

Isotr0py committed 59 days ago

Verified 55064354

[MRV2][KVConnector] Fix missing build_connector_worker_meta (#38698)

ivanium committed 59 days ago

Verified 311c9816

[CI/Build] Add audio deps in Dockerfile.cpu (#38876)

bigPYJ1151 committed 59 days ago

Verified 21d7ecc5

[Bug] Add e_score_correction_bias to SKIP_TENSORS (#38746)

hao-aaron committed 59 days ago

Verified 4729b908

full cudagraph for flex-attn (#36298)

shunting314 committed 59 days ago

Verified 8b141ed8

[Model] Add Phi4ForCausalLMV for microsoft/Phi-4-reasoning-vision-15B (#38306)

varun-sundar-rabindranath committed 59 days ago

Verified 2ad7c033

[CI][ROCm] Add Qwen3.5-35B-A3B-MXFP4 model eval into CI (#38664)

BowenBao committed 59 days ago

Verified 201d2ea5

[ROCm][Quantization][1/N] Refactor quark_moe w_mxfp4 w/ oracle (#38774)

BowenBao committed 59 days ago

Verified 103f0de5

refactor hard coded device string in test files under tests/v1 and tests/lora (#37566)

wincent8 committed 59 days ago

Verified 32e0c0bf

[Perf] Batch KV cache swap copies via cuMemcpyBatchAsync (#38460)

Etelis committed 59 days ago

Verified 4a06e124

[Kernel] Fuse FP8 output quantization into merge_attn_states (#36518)

carlyou committed 59 days ago

Verified 3bc2734d

[mla] Support fused FP8/NVFP4 output quantization in MLA attention (#35792) (#36205)

carlyou committed 59 days ago

Verified 1f5ec288

Older