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

COMMITS

develop
July 29, 2026
苍
[Operator Mechanism] fix zero-size contiguous strides (#79555)
苍天荒 committed
d68b054
July 27, 2026
A
support EchoMTP sync params and moment (#79542)
AlAuAu committed
4eda9f7
F
[Operator Mechanism] Clean up BLAS APIs and improve int64 safety (#79534)
feixi committed
b6603f7
L
update empty_strided for compat (#79540)
liuruyan committed
350f982
W
[API Compatibility] Support ChangePrefixMatcher for aligned nn classes -part (#79539)
Wenfei (Charles) Qi committed
8bc0a2b
苍
[Execute Infrastructure] 跳过 FP8 Tensor 的 NaN 和 Inf 检查 (#79547)
苍天荒 committed
a2bc39a
July 25, 2026
N
[CodeStyle] Bump ast-grep to v0.45.0 and Ruff to v0.16.0 (#79546)
Nyako Shigure committed
6e8183f
July 24, 2026
Z
【FSDP】Add FSDP in dynamic (#78577)
ZhenxingLi committed
f14c723
July 23, 2026
L
fix write overflow in unstack kernel (#79533)
LiYuRio committed
7c2af2c
X
Add muon_max_group_bytes to split large Muon groups and avoid OOM (#79537)
Xiangrui Yu committed
38621ba
Y
[AP] Reorganize code for matmul_pass. (#79535)
Yiqun Liu committed
16c629c
July 22, 2026
O
Use double precision for the norm_p value in norm reduce GPU kernel (#79497)
omoYang committed
08de6be
苍
[Operator Mechanism] Fix `rms_norm`/`layer_norm` from `auto`/`int` to `int64_t` (#79526)
苍天荒 committed
9471312
O
[Operator Mechanism] Fix p_norm_grad kernel precision compatibility for GPU (#79481)
omoYang committed
1dfe2f2
F
Add zero-class check for hard-label cross entropy (#79523)
feixi committed
1587050
F
[Operator Mechanism] Remove BLAS APIs only used on CPU (#79513)
feixi committed
1152678
W
[API Compatibility] enhance `paddle.enable_compat` -part (#79391)
Wenfei (Charles) Qi committed
718431c
Z
Improve VMM V2 remap recovery and OOM diagnostics (#79514)
Zhang Ting committed
ad1d2d4
N
🔥 chore(ci): remove model benchmark pipeline (#79531)
Nyako Shigure committed
3323d55
H
overlap优先于group call优化,当二者都被使用的时候 (#79520)
Haze188 灏喆 committed
f20c77b
July 21, 2026
G
[CI] retire TensorRT from default CI and remove Go TensorRT test (#79444)
gouzil committed
e10a7f1
H
dev muon sharding optimizer comm overlap (#79428)
Haze188 灏喆 committed
2556248
苍
[Operator Mechanism] Fix `index_select` empty-axis validation (#79509)
苍天荒 committed
748b0df
苍
[Operator Mechanism] Fix `embedding` invalid index handling (#79510)
苍天荒 committed
0d8eafe
July 20, 2026
N
[Build] Packaging vendored liblapack dylib on ARM64 platform (#79516)
Nyakku Shigure committed
f2396e7
B
align complex mul with torch (#79469)
Bohao Fan committed
dfa69d5
X
update group_NS (#79482)
Xiangrui Yu committed
80cc5d8
July 18, 2026
Z
[Memory] Support block info queries for VMM V2 (#79487)
Zhang Ting committed
49028ae
July 17, 2026
L
[Build] Avoid FlashAttention v3 for sm100 builds (#79500)
liuhao2638 committed
8d917cd
L
[User Experience] Muon optimizer supports lr_ratio (#79466)
LCStayingdullCircuit committed
5390960