Morph MORPH
SIGN IN SIGN UP
dhruvhead / Paddle UNCLAIMED
0 0 1 C++

COMMITS

develop
September 29, 2026
W
[CI] Remove Check file paths jobs (#79826)
Wang Shigure committed
4793e33
September 28, 2026
W
[CI] Remove pip cache from codestyle workflow (#79824)
Wang Shigure committed
2f7a3ff
S
[FlexCheckpoint] Avoid the full metadata load in the local-resume check (#79807)
SUN Dong committed
e9a8a18
September 24, 2026
O
[Big Tensor] Support iteration space beyond INT32_MAX in GPU index_elementwise_put kernels (#79810)
omoYang committed
67dafa8
O
Add kMaxBlockDimZ limit in GetGpuLaunchConfig3D to prevent block.z exceeds cuda limit of blockDim.z (#79808)
omoYang committed
ea6ba61
September 22, 2026
L
[compat] Resolve a scoped spec once instead of re-entering the finder (#79798)
LiYuRio committed
7c722e5
Y
代码缓存拉取submodule (#79801)
YUNSHEN XIE committed
54bbec0
F
[Operator Mechanism]Add CUDA small-stride index gradient kernel (#79716)
feixi committed
8b0a952
Y
增加Paddle代码缓存 (#79797)
YUNSHEN XIE committed
170026c
W
[FlexCheckpoint] Simplify AOA name mapping and trim the generation surface (#79791)
wangbingguang committed
6d1e60b
September 21, 2026
X
Translate Chinese comments to English in muon optimizer coverage test (#79794)
Xiangrui Yu committed
14d4108
S
[Performance] Add num_workers to paddle.load for parallel payload reading (#79786)
SUN Dong committed
92d36be
X
Support hyperball optim-method and shardingV1 for adam (#79792)
Xiangrui Yu committed
8a0db49
September 20, 2026
L
[RL FIX] Fix nopp rl error (#79680)
liuruyan committed
164bdb3
Y
Fix the bug in ShardingOptimizer when param and grad have different shape (#79788)
Yichen Zhang committed
3cd50a5
F
[Operator Mechanism] Add int64-safe batched GEMM and align bmm N==1 with Torch (#79772)
feixi committed
23f14e7
O
skip deterministic tests for take_along_axis_grad at ROCm (#79789)
omoYang committed
bda73da
September 17, 2026
正
Add machine-balanced partitioning for Muon 2D sharding (#79763)
正在学习 committed
01d68e3
O
Fix 0-D index bug and add deterministic path in gather grad kernel (#79754)
omoYang committed
5e9de01
O
[Operator Mechanism] Add deterministic path for take_along_axis_grad cuda kernel (#79753)
omoYang committed
4149585
September 16, 2026
L
[Performance] Parallelise moe_permute (#79756)
LiYuRio committed
13d53f4
September 15, 2026
R
删除icafe建卡检查 (#79771)
Ray961123 committed
c817668
B
align trunc_normal with torch for gpu (#79703)
Bohao Fan committed
e958ce4
G
dev commit (#79770)
Gu Shiwei committed
75405e9
F
Fix RestrictedUnpickler dotted-name whitelist bypass (#79764)
feixi committed
d80eb72
F
[Bug Fix] Align CPU elementwise broadcast backward with PyTorch bitwise (#79750)
feixi committed
7ec30a8
F
[Operator Mechanism] Consolidate strided view validation and gradient fixes (#79748)
feixi committed
96f4842
September 14, 2026
Z
[Accuracy] Align CUDA bilinear interpolation with deterministic Torch (#79762)
Zhan Rongrui committed
b61e01d
September 11, 2026
F
Align CPU var and std with PyTorch (#79747)
feixi committed
9a64753
September 10, 2026
L
[Execute Infrastructure] Rename ai_edited_test to coverage_test (#79738)
liuhao2638 committed
66ce0f3