strisunshinewentingwang/pytorch

Pull Requests Commits

fix where backward to use scalar 0 (#83043)

Natalia Gimelshein committed 2 years ago

e77d4ec5

[SSA] Clear Shape Cache on Load Error (#82850)

Gamrix committed 2 years ago

111214ac

[functorch] add error inputs check to vmap test (#83017)

zou3519 committed 2 years ago

24d3ea61

[functorch] better error when vmap over no Tensor inputs (#83016)

zou3519 committed 2 years ago

da8fcc42

[functorch] `squeeze_` batch rule, fix `squeeze_.dim` batch rule (#82972)

zou3519 committed 2 years ago

571b9301

Fix error_inputs for linalg.lstsq; assert SampleInput args are tuple (#83004)

zou3519 committed 2 years ago

21ec2884

[xla hash update] update the pinned xla hash (#83058)

pytorchmergebot committed 2 years ago

6c60a656

Use flatbuffer of alternate namespace (#82952)

qihqi committed 2 years ago

f9533560

Get rid of ENABLE_UPGRADERS macro (#77574)

tugsbayasgalan committed 2 years ago

b4b60c2a

[MPS] Make softplus beta a placeholder (#81169)

qqaatw committed 2 years ago

4a1783ee

Use UnaryUfuncInfo for type conversion functions (#82349)

peterbell10 committed 2 years ago

fc0d5d2b

Allow ufunc OpInfos to have no reference (#82348)

peterbell10 committed 2 years ago

5e3d1ef4

Fix prims.div to return the correct dtype (#82949)

peterbell10 committed 2 years ago

a17211f7

Fix IMA for topk (#83042)

zasdfgbnm committed 2 years ago

9e65e93c

Add nested tensor contiguous (#82147)

drisspg committed 2 years ago

e8166444

Supports symbolic diff for silu (#81724)

yueyericardo committed 2 years ago

09181549

Remove unnecessary copy constructor (#83030)

malfet committed 2 years ago

c6bc7661

Governance process been actualized. (#82736)

b0noI committed 2 years ago

9ba1631c

[ONNX] Reland #81953 Type utility for converting among JIT, torch and ONNX data types (#82995)

justinchuby committed 2 years ago

c6cdca5c

support custom class in torchgen schema parser (#82925)

shunting314 committed 2 years ago

94355396

Update cross entropy documentation to metion logits clearly (#82538)

Alex7Li committed 2 years ago

1fedd404

Revert "Update the pull request template (#81991)" (#83025)

malfet committed 2 years ago

dc51fc71

adding a custom caster for c10::SymInt (#82692)

Krovatkin committed 2 years ago

dee63f4f

remove unused/debug header (#82845)

Krovatkin committed 2 years ago

35b4ac4e

[PyTorch] Proper reset execution graph data in remove callback registration (#82910)

louisfeng committed 2 years ago

84079f31

[MPS] Remove checks that incur unnecessary syncs on GPU with tensor.item() (#82505)

razarmehr committed 2 years ago

8d1ff9fc

[BE] Add `append_cxx_flag_if_supported` macro (#82883)

malfet committed 2 years ago

d7e6aaa5

ci: Bump max_available for windows gpu to 100 (#83018)

seemethere committed 2 years ago

86d5262e

Fix typos in `torch.package` documentation (#82994)

bencwallace committed 2 years ago

7e3c3fd3

Allow map location to meta device (#82603)

ezyang committed 2 years ago

b9c8db43

Newer Older