strisunshinewentingwang/pytorch

Pull Requests Commits

[vision hash update] update the pinned vision hash (#85380)

pytorchmergebot committed 2 years ago

08f413bd

Address eellison's CR comments on AOTAutograd (#85370)

ezyang committed 2 years ago

75451d3c

[BE] Mark unused range-loop vars with `C10_UNUSED` (#85383)

malfet committed 2 years ago

3c870dad

Revert "Improve and expose cpp_backtrace to python binding (#84896)"

pytorchmergebot committed 2 years ago

3122a96e

Exposing native _scaled_dot_product_attention to torch.nn (#85044)

drisspg committed 2 years ago

9fdd8a8b

Remove TS compatibility transition code (#85003)

mikekgfb committed 2 years ago

d7029fea

Improve and expose cpp_backtrace to python binding (#84896)

SherlockNoMad committed 2 years ago

73fbca1e

Update torch.ops.aten.all ref to be symbolic-trace friendly (#85352)

wconstab committed 2 years ago

52fd7e49

[PyTorch] StorageImpl: cache size_bytes.is_symbolic() (#85309)

swolchok committed 2 years ago

f6a18d3d

Fixed memory issues in linalg_lstsq (#85357)

Chillee committed 2 years ago

90fa744c

fx quant: fix bug in custom module test (#85344)

vkuzo committed 2 years ago

cb8e73bb

Fix indentation in functorch limitations docs (#85346)

mplemay committed 2 years ago

62786a09

Setup fake tensor and symbolic shapes once at beginning of AOTAutograd (#85233)

ezyang committed 2 years ago

e1f63475

Correctly handle duplicate arguments to AOTAutograd (#85301)

ezyang committed 2 years ago

5f623f5c

Added `Tensor.to` overloads to `torch._refs.to` (#84802)

jjsjann123 committed 2 years ago

b9b27f76

[fix] composite compliance: cumprod, _masked.cumprod, linalg.vander (#85330)

kshitij12345 committed 2 years ago

d3dec809

Remove errant semicolon (#85356)

ezyang committed 2 years ago

e24e1791

Fake: fix conv_transpose2d striding (#82846)

eellison committed 2 years ago

a3afb2c2

[MPS] Handle reduction of scalars in edge-cases (#83743)

kulinseth committed 2 years ago

e1ed485c

Adding multigammaln ref and fix arange (#85153)

lezcano committed 2 years ago

d17b144e

[mta] APEX style Fused Adam (#81705)

crcrpar committed 2 years ago

7a6c4d0c

[pytorch] Inline std::forward definition (#85255)

VasuAgrawal committed 2 years ago

00a10652

Make ones and zeros's ref accepts variadic size argument (#85117)

SherlockNoMad committed 2 years ago

9c1a6a52

Fix fake_tensor to_copy meta dispatch (#85337)

wconstab committed 2 years ago

5e8f16b8

[CUBLAS][CUDA GRAPHS] (re-open of #83461) Explicitly set the workspace for cuBLAS handles (#85292)

eqy committed 2 years ago

4012e623

Add a reset() method to nvFuser FusionCache to enable proper resetting during tests. (#85319)

kevinstephano committed 2 years ago

39f482ac

Revert D39583438: Multisect successfully blamed D39583438 for test or build failures (#85277)

benoitsteiner committed 2 years ago

86d8c61c

Add __all__ to torch.{fx, distributed, backends} submodules (#85079)

anjali411 committed 2 years ago

cf2f552c

[composite compliance] prod (#81969)

kshitij12345 committed 2 years ago

a4dca982

[MPS] Fix conv1d backwards crash for channels last case (#85283)

kulinseth committed 2 years ago

077db3de

Newer Older