strisunshinewentingwang/pytorch

Pull Requests Commits

[BE] Revert distributed change in https://github.com/pytorch/pytorch/pull/68779 (#83181)

rohan-varma committed 2 years ago

b29a0748

[FSDP] Remove unneeded checks (#83150)

rohan-varma committed 2 years ago

4e90526a

hash update - bug fix for branches (#83865)

clee2000 committed 2 years ago

8e074f45

[MPS] Move mps_linear to mps dispatch key (#80068)

drisspg committed 2 years ago

7cfc8b78

[cmake] Change COLORIZE_OUTPUT option to USE_COLORIZE_OUTPUT (#83716)

xwang233 committed 2 years ago

b18f9843

Lazy deprecation import function in torch.nn (#83834)

z-a-f committed 2 years ago

b7afee8a

fix upsample bf16 issue for channels last path by using high pricsion to compute index (#83847)

XiaobingSuper committed 2 years ago

658f958b

[ONNX] Add quantization support to more single output ops (#83008)

justinchuby committed 2 years ago

80cfafc3

Add lazy shape inference for cholesky op (#83720)

wonjoolee95 committed 2 years ago

1e4383f7

Migrate last workflows from 18.04 to 22.04 (#83861)

janeyx99 committed 2 years ago

36f6d91a

[optim] rmsprop: handle complex params as independent real params (#83860)

kshitij12345 committed 2 years ago

09331c94

Fix model type CNN->MLP in functorch ensembling notebook intro (#83603)

janosh committed 2 years ago

62d9f155

Used generator for "any" and "all" (#83844)

rezasharifi committed 2 years ago

dc557b94

[maskedtensor] adding unary and binary operations (#82837)

george-qi committed 2 years ago

e10c47a7

[ONNX] Introduce ONNXScopeName (#82038)

BowenBao committed 2 years ago

daca0ee5

[mergebot] Post PR Comment on cancel (#82744)

zengk95 committed 2 years ago

91766360

More doctest refinements. (#83317)

Erotemic committed 2 years ago

b136f3f3

modify the signature of method `__getitem__` from `ModuleList` (#83799)

Fabian Ricardo Latorre Gomez committed 2 years ago

9c9f4248

Move _masked opinfos to opinfo/definitions/_masked.py (#83763)

peterbell10 committed 2 years ago

91eb1b9b

Move `torch.special` OpInfos into opinfo/definitions/special.py (#83762)

peterbell10 committed 2 years ago

7656ef73

Fix for transposed convolution shape functions (#83557)

gpetters94 committed 2 years ago

35d4fa44

[JIT SSA] Allow updating shape functions without recompilation (#83629)

Gamrix committed 2 years ago

eff28d61

Revert "Optimize transpose copy on CPU using fbgemm transpose (#83327)"

pytorchmergebot committed 2 years ago

53cda905

[MPS] Fix the index error in constant_pad_nd() with single-dimension input (#83745)

razarmehr committed 2 years ago

d1be36ce

[MPS] Fix placeholder case for missing gather graph (#83744)

DenisVieriu97 committed 2 years ago

a6b75bb0

[Quant] Separate FBGEMM/QNNPACK BackendConfigs (#83566)

andrewor14 committed 2 years ago

b8496eb4

make sym sizes be computed lazily (#82233)

Krovatkin committed 2 years ago

07d0c9ec

Optimize transpose copy on CPU using fbgemm transpose (#83327)

CaoE committed 2 years ago

f56720ea

release the current symintnode in the move c-tor (#83789)

Krovatkin committed 2 years ago

fcb12440

Fix uninitialized member if the default c-tor is called (#83788)

Krovatkin committed 2 years ago

b47f712b

Newer Older