-
Notifications
You must be signed in to change notification settings - Fork 713
Pull requests: vllm-project/vllm-ascend
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[BugFix] NetLoader: No backend type associated with device type npu
#5700
opened Jan 7, 2026 by
destinysky
Loading…
[DRAFT] [FIX] Replace the implementations of o_proj, q_b_proj, and kv_b_proj with custom_op for sharded CP
documentation
Improvements or additions to documentation
merge-conflicts
module:core
module:ops
module:tests
#5698
opened Jan 7, 2026 by
zzhx1
Loading…
[doc](pcp) correct the seq length of KV for prefill of GQA
documentation
Improvements or additions to documentation
#5697
opened Jan 7, 2026 by
pisceskkk
Loading…
Optimize the print info format when deprecated code is used in vllm-ascend
module:core
#5696
opened Jan 7, 2026 by
leo-pony
Loading…
[BugFix][P/D] Fix pre-create link parameter error
ready
read for review
ready-for-test
start test by label for PR
#5694
opened Jan 7, 2026 by
nwpu-zxr
Loading…
[main][bugfix] Fix fullgraph padding bug in mtp eagle refactor
#5692
opened Jan 7, 2026 by
lilinsiman
Loading…
[Main2Main] Upgrade vllm commit to 0107
ci/build
documentation
Improvements or additions to documentation
module:tests
ready
read for review
ready-for-test
start test by label for PR
#5691
opened Jan 7, 2026 by
zhangxinyuehfad
Loading…
[Verify] Upgrade CANN 8.5.0.alpha002
ci/build
documentation
Improvements or additions to documentation
module:tests
#5688
opened Jan 7, 2026 by
wjunLu
Loading…
fix 310p unable to run Qwen 2.5dense/vl and Qwen3 dense models
merge-conflicts
module:core
module:ops
#5686
opened Jan 7, 2026 by
Tflowers-0129
Loading…
[bugfix]support dsv3.2 enable both mtp and full_decode_only
ready
read for review
ready-for-test
start test by label for PR
#5679
opened Jan 7, 2026 by
cookieyyds
Loading…
Eliminate H2D copy bubbles by leveraging asynchronous stream scheduling.
documentation
Improvements or additions to documentation
module:tests
#5677
opened Jan 7, 2026 by
mengxingkongzhouhan
Loading…
[Bugfix] fix the number of batch is incorrect in Xlite when sequence …
#5675
opened Jan 7, 2026 by
changdawei1
•
Draft
Ensure that the qwen3-vl model uses the mrope operator(fix issue#5670)
module:ops
#5673
opened Jan 6, 2026 by
Xiangyanglikecode
Loading…
[feature]dcp&pcp support mlapo
ready
read for review
ready-for-test
start test by label for PR
#5672
opened Jan 6, 2026 by
zhenwenqi2024
Loading…
VL model enable flashcomm V1
merge-conflicts
module:core
module:ops
#5669
opened Jan 6, 2026 by
shaopeng-666
Loading…
Add Medusa speculative decoding support for vllm_ascend
#5668
opened Jan 6, 2026 by
simplzyu
Loading…
support TensorList for dispatchFFNCombine
ready
read for review
ready-for-test
start test by label for PR
#5665
opened Jan 6, 2026 by
lhchg
Loading…
[Ops] replace _update_out_and_lse with _npu_attn_out_lse_update
module:tests
#5662
opened Jan 6, 2026 by
YzTongNiar
Loading…
[Feat] Remove Redundant Variables after Integrate FIA operator in mla_cp._forward_decode
module:tests
#5659
opened Jan 6, 2026 by
dsxsteven
Loading…
Previous Next
ProTip!
Updated in the last three days: updated:>2026-01-04.