strisunshinewentingwang/pytorch

Pull Requests Commits

Add Meta backend

ezyang committed 2 years ago

5ca24c60

Port linalg_qr to structured

lezcano committed 2 years ago

af6321f3

[quant][core][feature] Implemented masked_fill for QuantizedCPU tensors

dzdang committed 2 years ago

2b6a0427

Run torchdynamo tests on PyTorch Linux CI

ezyang committed 2 years ago

fe2fbb94

Revert "Added {logical_not, trace} refs, moved logical ops to use method overloads"

pytorchmergebot committed 2 years ago

d67309ae

Revert "Added kl_div_backward decomp"

pytorchmergebot committed 2 years ago

60a13f4e

Print output during MPS test import tests (#79163)

malfet committed 2 years ago

97594a24

Revert "`kl_div`: fix for grads wrt `target`, double backward, forward-over-reverse AD support. (#79007)"

pytorchmergebot committed 2 years ago

3556457d

Revert "formatted _decomp folder with black"

pytorchmergebot committed 2 years ago

2027eae6

MAINT: Harmonize argsort params with array_api (#75162)

HaoZeke committed 2 years ago

04407431

Support CSC/BSR/BSC inputs to unary zero-preserving functions.

pearu committed 2 years ago

fb6749d9

Support ComplexHalf in nonzero and add of sparse_csr input.

pearu committed 2 years ago

2a0e4322

`kl_div`: fix for grads wrt `target`, double backward, forward-over-reverse AD support. (#79007)

nikitaved committed 2 years ago

72ad222c

Accept non-standard bools in more CUDA kernels

peterbell10 committed 2 years ago

cd9e1580

formatted _decomp folder with black

Chillee committed 2 years ago

4945c721

Added kl_div_backward decomp

Chillee committed 2 years ago

a08685eb

Added {logical_not, trace} refs, moved logical ops to use method overloads

Chillee committed 2 years ago

64b6bd8c

Minor fix in FX tests with TorchDynamo (#79174)

anijain2305 committed 2 years ago

28ed0f56

Revert "[mobile] Fix lightweight dispatch OOM error by introducing selective build"

pytorchmergebot committed 2 years ago

c3e089a0

[ci] don't upload test stats for cancelled workflows

suo committed 2 years ago

cb04cc0a

Apply clang-format to ATen headers (#78406)

malfet committed 2 years ago

95b15c26

[vulkan] Implement replication_pad2d op (#79057)

kirklandsign committed 2 years ago

5ea3e657

Revert "Support both train / eval modes for ModuleInfo"

pytorchmergebot committed 2 years ago

854c833f

Improved stack ref and added more decomposition annotations

Chillee committed 2 years ago

dc11a564

Ported roll to use torch ops and added as a decomposition

Chillee committed 2 years ago

c3531c9b

Add PReLU to MKLDNN convertible Ops (#79011)

xsacha committed 2 years ago

1b75a2e4

Black torch._meta_registrations

ezyang committed 2 years ago

225bf132

Do not treat all dense tensors as isTensorSubclassLike

ezyang committed 2 years ago

eb856daf

Add vision onto the commit pins train

ezyang committed 2 years ago

8cd23eb4

Centralize commit pins in a folder

ezyang committed 2 years ago

004b6387

Newer Older