strisunshinewentingwang/pytorch

Pull Requests Commits

Use 12xlarge for nightly cpp doc generation job (#86859)

huydhn committed 2 years ago

5acf6e0e

[dynamo] Introduce `get_real_value` API to TensorVariable (#87091)

suo committed 2 years ago

48142707

[docs] Fix ScalarTensor __repr__ in Extending PyTorch example (#86330)

jmargeta committed 2 years ago

e85dbcc9

[Dynamo] More robust pyop support, module properties as args (#87020)

voznesenskym committed 2 years ago

b8007742

[ONNX] Ignore print(Tensor) during tracing (#86223)

Thiago Crepaldi committed 2 years ago

1167949b

Workarounds for cudnn_batch_norm with TorchRefsNvfuserCapabilityMode (#86796)

IvanYashchuk committed 2 years ago

31931515

add XLA backend into tensor type strings (#86881)

holimion committed 2 years ago

33343def

Revert "OpInfo: Sample input cleanup (4/n) (#86324)"

pytorchmergebot committed 2 years ago

317eeb81

Give more clear error message when gscope is non-empty (#87005)

JackCaoG committed 2 years ago

8f85831f

[DataPipe] Fix missing functional name for FileLister (#86497)

NivekT committed 2 years ago

c01c7a5e

Update action lint with missing new runners from scale-config (#87009)

huydhn committed 2 years ago

c27a5171

Enables `where` to have cpu scalar args (#87022)

Natalia Gimelshein committed 2 years ago

1704256b

[functorch] Fix cross to match unbatched behavior (#86926)

samdow committed 2 years ago

f3969bd8

Avoid calling logging.basicConfig (#86959)

SherlockNoMad committed 2 years ago

e271e823

Add meta support for _adaptive_avg_pool2d_backward (#86359) (#87074)

anjali411 committed 2 years ago

63512205

Revert "[Dynamo] More robust pyop support, module properties as args (#87020)"

pytorchmergebot committed 2 years ago

66715767

fix cudagraphify for inplace parameter change (#87060)

Natalia Gimelshein committed 2 years ago

8617f5f4

Revert "[inductor] Use decomps for unfold (#87025)"

pytorchmergebot committed 2 years ago

2c6167c4

[inductor] Set correct strides in fallback example run (#87049)

anijain2305 committed 2 years ago

2b558138

ATen/native (2/6): Use per-operator headers (#75572)

peterbell10 committed 2 years ago

4e5357fa

conv backward impl (#87047)

albanD committed 2 years ago

b40f4434

autograd clone_obey_contract() symint support (#87044)

albanD committed 2 years ago

1463013c

meta funcs for avg_pool2d and avg_pool2d_backward (#87043)

albanD committed 2 years ago

86c2e44c

Very limited pow support (#87042)

albanD committed 2 years ago

c21dcffc

[xla hash update] update the pinned xla hash (#87067)

pytorchmergebot committed 2 years ago

37e9e89a

[primTorch] Add a ref for `narrow_copy` (#86748)

nkaretnikov committed 2 years ago

91b3cd0b

[primTorch] Implement NLL loss reference (#81128)

rdspring1 committed 2 years ago

847ded6d

[TorchInductor] enable inplace buffers by default (#87037)

jgong5 committed 2 years ago

78e22890

Make `AdamW`, `NAdam` & `RAdam` differentiable (#86183)

Emilio Castillo committed 2 years ago

1b43883f

[vision hash update] update the pinned vision hash (#87021)

pytorchmergebot committed 2 years ago

364a9973

Newer Older