strisunshinewentingwang/pytorch

Pull Requests Commits

Fix indentation in functorch limitations docs (#85346)

mplemay committed 1 year ago

62786a09

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

ezyang committed 1 year ago

e1f63475

Correctly handle duplicate arguments to AOTAutograd (#85301)

ezyang committed 1 year ago

5f623f5c

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

jjsjann123 committed 1 year ago

b9b27f76

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

kshitij12345 committed 1 year ago

d3dec809

Remove errant semicolon (#85356)

ezyang committed 1 year ago

e24e1791

Fake: fix conv_transpose2d striding (#82846)

eellison committed 1 year ago

a3afb2c2

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

kulinseth committed 1 year ago

e1ed485c

Adding multigammaln ref and fix arange (#85153)

lezcano committed 1 year ago

d17b144e

[mta] APEX style Fused Adam (#81705)

crcrpar committed 1 year ago

7a6c4d0c

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

VasuAgrawal committed 1 year ago

00a10652

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

SherlockNoMad committed 1 year ago

9c1a6a52

Fix fake_tensor to_copy meta dispatch (#85337)

wconstab committed 1 year ago

5e8f16b8

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

eqy committed 1 year ago

4012e623

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

kevinstephano committed 1 year ago

39f482ac

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

benoitsteiner committed 1 year ago

86d8c61c

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

anjali411 committed 1 year ago

cf2f552c

[composite compliance] prod (#81969)

kshitij12345 committed 1 year ago

a4dca982

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

kulinseth committed 1 year ago

077db3de

[MPS] Fix the crash in bitwise ops on x86 platforms. (#85285)

kulinseth committed 1 year ago

bcdef58a

[Quant] Improve performance of ONEDNN backend (#84470)

Xia-Weiwen committed 1 year ago

6c48a01c

Revert "Python stack tracing OD flow (part 1) (#84362)"

pytorchmergebot committed 1 year ago

35088f28

[vision hash update] update the pinned vision hash (#85315)

pytorchmergebot committed 1 year ago

8c7e2097

[torch.futures] Fix nullptr deref (#85304)

malfet committed 2 years ago

c05ca0db

[functorch] Fix dangling impls (#85299)

zou3519 committed 2 years ago

66907e72

Revert "Reduce memory usage requirement of `test_warp_softmax_64bit_indexing` in `test_nn.py` (#85037)"

pytorchmergebot committed 2 years ago

53fdd606

Fix segfault for `out` with a large number of dims (#85294)

ngimel committed 2 years ago

a998a8eb

Setup torch/csrc/functorch/*; move CompileCache.{h, cpp} there (#85263)

zou3519 committed 2 years ago

d9024ea2

Python stack tracing OD flow (part 1) (#84362)

slgong-fb committed 2 years ago

1f4f05e5

Reduce memory usage requirement of `test_warp_softmax_64bit_indexing` in `test_nn.py` (#85037)

eqy committed 2 years ago

66a9cba2

Newer Older