COMMITS
April 16, 2026
Q
[fix]: align _flat_weights during functional calls
qq_35274980 committed
K
fix transfer_to_npu has_triton patch bug
kkjocker committed
April 15, 2026
P
Update op_plugin commit id
pta-robot committed
T
[fix] Add ephemeral timeout support for collective operations
TrHan committed
M
[refactor]添加sk options
mihudan committed
L
[sync] PR-33676: 【fix】ut of _host_emptyCache patch
LiNuohang committed
P
Update torchair commit id
pta-robot committed
W
[fix] fix found_inf param in async allreduce
wanlinan committed
H
Fix precision issue in MLIR mode fallback op execution
HandsoemLemon committed
April 14, 2026
P
Update op_plugin commit id
pta-robot committed
L
Y
[fix] ascend_launch_blocking on replay, add sync
yurongkun committed
S
rewrite_dropout_meta_and_decompose
sunhaolun committed
B
add torch mlir ut case
bigprestigee1 committed
H
[fix] fix the bug of FSDPTestMultiThread's _world install
hanye02 committed
L
fix: fix from blob bug
luochao60 committed
April 13, 2026
P
Update op_plugin commit id
pta-robot committed
P
Update op_plugin commit id
pta-robot committed
P
Update torchair commit id
pta-robot committed
G
del 32 padding size with cann version check
glf2016 committed
Z
fixing review comments on master
zhangguoguang committed
W
fix: handle zero tensor src to avoid null pointer in aclrtMemcpy
weixin_47897441 committed
G
npugraphex config update
glf2016 committed
P
Update torchair commit id
pta-robot committed
Y
【fix】transfer_to_npu adapt pg_nccl
yule100 committed
April 11, 2026
P
Update op_plugin commit id
pta-robot committed
April 10, 2026
L
[fix] patch torch._C._host_emptyCache
LiNuohang committed
Y
feat: adapt deterministic auto-tuning for triton and mlir
yvjc committed
P
Update op_plugin commit id
pta-robot committed