Morph MORPH
SIGN IN SIGN UP
dhruvhead / vllm-ascend UNCLAIMED
0 0 107 C++

COMMITS

main
August 27, 2026
W
[Test][BugFix] Remove invalid finegrained_tp_config from Kimi-K2.6 test config (#15065)
wangzhishenghw committed
ff450cb
L
[CI] fix nightly bug from sp enabled (#15017)
lilinsiman committed
7702ccd
X
[Feature][MRV2] Support lmhead TP in model runner v2 sample path (#14668)
XuChi committed
bed43af
W
[CI] Add .dockerignore and increase pip --timeout to 1200s (#14964)
whjnbm committed
76421f1
K
[Refactor][Config] Remove msMonitor daemon environment variable (#14828)
kk_1999 committed
f8958a8
K
[Refactor][Config] Remove balance scheduling environment variable (#14827)
kk_1999 committed
9bef60a
August 26, 2026
G
[CI]fix the failed weekly cases (#14972)
guxin108 committed
e8d9c1a
J
[BugFix][SpecDecode] Pad draft sequence lengths for full graph (#14801)
jiajinzhu2 committed
f9945a8
R
[Feature][MRV2] Support DSA decode ACLGraph (#14762)
RenYuKai committed
40f94b7
R
[BugFix][Config] Disable static kernel for MiniMax tests (#15034)
RuiZhang committed
aa8ca8b
L
[Refactor]unify calls to torch_npu.npu_mla_prolog_v3 (#14676)
luomin2005 committed
128ecf8
S
[CI] Overhaul PR test scheduling with shared 800I/800T NPU pools (#14793)
SILONG ZENG committed
b7fc48b
X
[CI][BugFix][MRV2]Down the accptance of dspark (#14890)
Xiaoshuang Wang committed
1d9cf24
S
[CI][Misc] Remove the gitleaks binary, add it to .gitignore (#15010)
Sijie Fu committed
48f4fa5
J
[CI][MRV2] Temporarily skip DeepSeek V4 MTP eager test (#15007)
jiajinzhu2 committed
ce8f64e
A
[Test][Feature] Add Qwen3.5-27B multi-LoRA MTP eager serving E2E (#14794)
Agoni-02 committed
9c7ef37
K
[Refactor][Config] Remove fused MC2 environment variable (#14831)
kk_1999 committed
80c833f
Z
[CI] Fix run-selected-tests-a5 skipped under ready-all label (#14789)
zhangxinyuehfad committed
e8600b7
Z
[CI] Restructure model download list by cluster and update workflow (#14908)
zhangxinyuehfad committed
633ab8e
Z
[CI] Add A5/950 wheel build job (#13287)
zhangxinyuehfad committed
92b0187
W
[BugFix]fix sfa + dcp cp_kv_cache_interleave_size setting (#14876)
weiguihua2 committed
1d81c26
S
[Feature][310P] Adapting MRV2 for the 310P (#14761)
SHUAI YANG committed
59a36f2
Z
[CI] Add cut branch checklist issue template (#13253)
zhangxinyuehfad committed
9022479
J
[Performance] use sort-based top-k/top-p in non-reduce mode for consistent performance (#13842)
JC-ut0 committed
994765d
X
[BugFix] Add mlapo_keep_prefill_weights option to keep prefill weights on NPU (#14216)
xlshaoscu committed
ca56038
L
[BugFix] Filter target-only attention layers from draft prefill metadata (#14791)
Liao Qingqing committed
895d475
Y
[Feature][Quantization] Serve native block-wise FP8 checkpoints for qwen3.8-27b-fp8 (#14852)
yiminghub2024 committed
061a926
A
[Feature][MRV2][Bugfix] Add a triton ops patch to avoid failure from upstream (#14862)
AuroraEmiya committed
e5365a6
L
[Test] Remove obsolete MRV2 slot-mapping test (#14893)
leolee committed
ff998aa
L
[Doc][MRV2] Add prefill context parallelism guide (#14571)
leolee committed
2452b1f