strisunshinewentingwang/pytorch

Pull Requests Commits

BatchNorm: fix mixed precision usage with affine=False (#61962)

peterbell10 committed 3 years ago

0df1679e

Ignore LNK4099 for debug binary libtorch builds (#62060)

janeyx99 committed 3 years ago

e318058f

ns for fx: expose hook to define custom weight extraction functions (#62047)

vkuzo committed 3 years ago

04c95a06

ns for fx: fix typing issue in weight extraction (#62041)

vkuzo committed 3 years ago

07c6a120

ns for fx: change weight extraction to direct mapping (#62038)

vkuzo committed 3 years ago

eaba16d6

Fix some sign comparisons (#61849)

r-barnes committed 3 years ago

8a2c525d

Migrate scheduled jobs debuggability to GHA (#62056)

janeyx99 committed 3 years ago

9d405646

[DDP Comm Hook] Use a single tensor instead of a tensor list as the comm hook result (#62074)

Yi Wang committed 3 years ago

b03b45af

`clamp`: port to structured kernel (#61361)

Meghan Lele committed 3 years ago

1d2ea76a

preserve residual in transformer norm_first (#61692)

Basil Hosmer committed 3 years ago

b106b958

Reformat (#62073)

Yi Wang committed 3 years ago

53222c59

[pruner] add Conv2d support (#61778)

kazhou committed 3 years ago

3687bbb1

Disable `avoid-non-const-global-variables` lint check (#62008)

malfet committed 3 years ago

a9b0a921

Disable bazel in CircleCI (#62055)

malfet committed 3 years ago

260198d4

Modernize make pointers (#61741)

r-barnes committed 3 years ago

a91be24e

zasdfgbnm committed 3 years ago

c998f2d9

zasdfgbnm committed 3 years ago

611bfb0b

[skip ci] minor typo link fix (#62042)

janeyx99 committed 3 years ago

f98fa5ea

.github: Only run workflows on pytorch/pytorch (#62044)

seemethere committed 3 years ago

1a64a5c0

zasdfgbnm committed 3 years ago

3ddf67cf

DOC Fixes link in register_module_backward_hook (#61999)

thomasjpfan committed 3 years ago

414537ac

zasdfgbnm committed 3 years ago

244ce350

Svd docfix (#62028)

albanD committed 3 years ago

b522f3be

zasdfgbnm committed 3 years ago

621b5aba

Add build/.ninja_log to artifacts for Windows (#62035)

janeyx99 committed 3 years ago

d6e776d9

zasdfgbnm committed 3 years ago

5911efba

ENH Adds no batch dim support for AvgPool1d (#61860)

thomasjpfan committed 3 years ago

0309c578

zasdfgbnm committed 3 years ago

c20575c5

Warn about poor performance creating Tensor from list of numpy.array's (#51680)

kurtamohler committed 3 years ago

5a00152a

[Static Runtime] Implement prim::isinstance and prim::TypeCheck (#61783)

Mike Iovine committed 3 years ago

2b0eddb0

Newer Older