vllm-project/vllm

Pull Requests Commits

[ROCm][CI] Keep startup profiling from aborting when free memory grows (#53591)

stefankoncarevic committed 25 days ago

Verified 9236159b

[Test] Fix mock_current_vllm_config missing kernel_config in test_dflash2 (#54132)

mayuyuace committed 25 days ago

Verified c01b50e3

[CPU][MLA] Fix prefill backend selection so MLA runs end-to-end on CPU (#51471)

maobaolong committed 25 days ago

Verified 6f7df92a

Count store offers, not lookups, for CPU offload store_threshold (#52227)

okorzh-amd committed 25 days ago

Verified e6bfe03a

[Bugfix][KV Cache] Prevent negative external block allocation (#52707)

haic0 committed 25 days ago

Verified 53bf9905

[Bugfix] Allocate packed outputs in fused_q_kv_rmsnorm so q_b_proj keeps its low-latency GEMM path at decode (#53109)

esmeetu committed 25 days ago

Verified b74ac0bc

Remove wrongly added e2e test (#54099)

elvircrn committed 25 days ago

Verified f3b91f1a

[Misc] Separate adaptive verification config validation (#54108)

njhill committed 25 days ago

Verified 8cfba696

[Bugfix] Remove race in fused groupwise RMSNorm quantization (#54111)

mgoin committed 25 days ago

Verified 5bbc58c0

[Kimi-K3] Merge MLA gate into QKV-A projection (#54015)

gau-nernst committed 25 days ago

Verified 6ec92bcb

fix(build): correct preprocessor guard for GDN decode to fix Ampere c… (#52743)

prakharPant committed 25 days ago

Verified b2a6e9dc

[Bugfix] Revert renderer warmup overlap to avoid fork deadlock (#54023)

khluu committed 25 days ago

Verified d04d3e0a

[Rust Frontend][gRPC] Enforce LoRA path validation across transports (#53756)

Connor Carpenter committed 25 days ago

Verified 3cc015ac

[Bugfix][Model] Fix K3 DSpark config for 96-head drafts (#54005)

zixi-qi committed 25 days ago

Verified e5377e67

[kernel] Integrate FlashInfer BF16 CuTeDSL Low Latency GEMM (#50572)

jiahanc committed 25 days ago

Verified 7c877062

Add support for loading dflash2 model in speculators format (#53797)

fynnsu committed 25 days ago

Verified f79a2f55

[Bugfix][MRV2] Isolate sleep-mode KV allocations (#53508)

Ronald1995 committed 25 days ago

Verified 479eeb32

Upgrade tpu-inference to v0.28.0 (#54020)

meiyeh123 committed 25 days ago

Verified d262964e

[Attention] Enable masked MHA for GLM-5 head dimensions (#53785)

MatthewBonanni committed 25 days ago

Verified de9250ac

[Bugfix][Parser] Scope reasoning-end detection to the current turn via turn-boundary tokens (#54089)

Xarbirus committed 25 days ago

Verified 3a9bfc20

[Perf][DSv4] Use native CUDA SwiGLU clamp kernel for Humming MoE (throughput +1.40%) (#53685)

chaunceyjiang committed 25 days ago

Verified 7d5769b2

[Kimi Perf] Tune hopper low latency gemm kernel, 4%~97% performance improvement (#54088)

yewentao256 committed 25 days ago

Verified 9818bb3d

[Bugfix] Avoid TCPStore port collision for co-located non-DP Ray engines (#53666)

jeffreywang88 committed 25 days ago

Verified 79378fe6

[Rust Frontend] Align OpenAI request and response edge cases (#53218)

BugenZhao committed 26 days ago

Verified 9db222c6

[Model Runner V2] Use MRV2 for all models by default (#53183)

njhill committed 26 days ago

Verified 4aab2b0e

[Doc] Add EXAONE-4.0-1.2B to batch invariance tested models (#53839)

cogniera committed 26 days ago

Verified a18dbe49

Fix Humming MoE activation_output aliasing (#54056)

elvircrn committed 26 days ago

Verified 674c284d

[Bugfix] Handle malformed namespace tools (#53763)

taneem-ibrahim committed 26 days ago

Verified d7b6a359

[Attention][DCP] Use FlashInfer native CP for MLA decode (#54012)

GirasoleY committed 26 days ago

Verified 07242faa

[ROCm][Perf] Fuse SWA q/kv RMSNorm and q FP8 group quant for DeepSeek-V4 (#53540)

shen-shanshan committed 26 days ago

Verified 32ad1400

Older