Morph MORPH
SIGN IN SIGN UP
tejas / vllm

A high-throughput and memory-efficient inference and serving engine for LLMs

0 0 122 Python

COMMITS

main
October 22, 2025
J
[Core] Handle MoE LoRA edge cases (#27335)
Jee Jee Li committed
abf3db4
G
Bugfix - pass 'max_num_tokens_padded' into 'moe_lora_align_block_size' (#27311)
gnovack committed
8e4ca4d
W
[Log] Add Warning for `LLM(data_parallel_size=k)` single-process DP Usage (#27282)
Wentao Ye committed
1a0f4de
L
[Bugfix][CPU] Disable dual stream execution for experts on CPU (#27320)
Li, Jiang committed
843af7f
W
[Frontend][3/N] Improve all pooling task | Support binary embedding response (#27066)
wang.yuqi committed
1f633b8
E
fixed reasoning streaming with tool_choice="required" (#24108)
ExtReMLapin committed
a4c29e6
H
Remove last `level` references not removed in #26355 (#27260)
Harry Mellor committed
8f18feb
H
Update release pipeline for PyTorch 2.9.0 (#27303)
Huy Do committed
ed540d6
W
[1/N][Platform] Cleanup useless function (#26982)
wangxiyuan committed
f6027b2
J
[torch.compile] Enable silu_mul_fp8_quant fusion without custom ops enabled (#27146)
Jiangyun Zhu committed
ab3e800
C
[Benchmark] Add plot utility for parameter sweep (#27168)
Cyrus Leung committed
ceacedc
N
[CI] Nixl integration tests DP-EP (#27199)
Nicolò Lucchesi committed
bfa59be
V
[DOC] [ROCm] Add ROCm quickstart guide (#26505)
vllmellm committed
265ecb0
October 21, 2025
L
[Deepseek v3.2] Remove extra logics in indexer (#26465)
Lain committed
09a7e6f
T
[P/D] KVConnector for decode benchmarking (#25986)
Tyler Michael Smith committed
6c2eef5
B
[Bugfix] skip cuda graph for drafter when running with eager (#26821)
Benjamin Chislett committed
1974880
E
Updated xgrammar backend to not deny supported string formats (#27253)
ExtReMLapin committed
4a8a567
A
[Performance] Dual stream execution of "shared_experts" and "selected_experts" inside FusedMoE (#26440)
Alexander Matveev committed
344a001
H
Update PyTorch to 2.9.0+cu129 (#24994)
Huy Do committed
becb7de
T
[Bugfix] fixes the decoding metadata of dense mla's fp8 kvcache. (#27144)
Tao He committed
250fb1b
N
[V0 Deprecation] Remove V0 executors (#27142)
Nick Hill committed
647214f
D
[Bugfix][P/D] Reduce num_threads used by nixl ucx backend (#27196)
David Whyte-Gray committed
ddeec11
W
[Feature] Batch Invariant for R1 TP 8 on Blackwell (#27229)
Wentao Ye committed
86ed770
M
[ROCm] Update Triton, Torch, and AITER branches for ROCm base Dockerfile (#27206)
Micah Williamson committed
aa1356e
P
Add @pavanimajety to .github/codeowners for Flashinfer, ModelOpt related code (#27213)
Pavani Majety committed
ecc3c09
J
[ROCM] Enable CompressedTensorsWNA16 (#27187)
JartX committed
ba09652
H
[CI] Install pre-release version of `apache-tvm-ffi` for `flashinfer` (#27262)
Harry Mellor committed
bd66b85
D
[Chore] Separate out NCCL utilities from vllm.utils (#27197)
dongbo910220 committed
6c728f7
D
[Deepseek v3.2] Optimize top_k_per_row (#26763)
Daniel Cámpora committed
80e9452
R
[MM][Core] Decouple ViT backend from LM backend (#27061)
Roger Wang committed
c3a2c6a