strisunshinewentingwang/pytorch

Pull Requests Commits

kwen2501 committed 2 years ago

f53036ba

kwen2501 committed 2 years ago

7c9d8b2d

Move getNcclErrorDetailStr into cpp

kwen2501 committed 2 years ago

531df69e

kwen2501 committed 2 years ago

d4defad1

Adopt ncclRemoteError

kwen2501 committed 2 years ago

1271a987

Add reduce_scatter_tensor in place of _reduce_scatter_base (#85867)

kwen2501 committed 2 years ago

ade1c196

[ONNX] Rename 'sarif_om' to 'sarif' (#85918)

BowenBao committed 2 years ago

33401ee8

[ONNX] Add type annotation for SARIF attributes (#85898)

BowenBao committed 2 years ago

6bb0a36d

[ONNX] Migrate SARIF from attr to dataclasses (#85651)

BowenBao committed 2 years ago

e9b254a0

[ONNX] Introduce SARIF (#85428)

BowenBao committed 2 years ago

91667d1d

Refactor distribuetd to use absolute header path (#85780)

minsii committed 2 years ago

1ad0048b

[Profiler] Extend ID assignment to allocations and frees (#85719)

Taylor Robie committed 2 years ago

0b0ce72b

Hotfix for S298125 (#85814)

ezyang committed 2 years ago

95681929

Revert "Refactor distribuetd to use absolute header path (#85780)"

pytorchmergebot committed 2 years ago

a50d8864

Refactor distribuetd to use absolute header path (#85780)

minsii committed 2 years ago

66808271

[MPS] Handle scalar input for scatter and gather (#85842)

abhudev committed 2 years ago

81b366a9

[MPS] Handle output shape for empty input in binary ops (#85836)

abhudev committed 2 years ago

62a4fd79

Make launch check exit code depend on results (#85886)

r-barnes committed 2 years ago

ae93a4dc

[c10d] Reorder macros so they are defined before getting used (#85850)

drej82 committed 2 years ago

dde43d08

[ONNX] Remove protocol dataclass (#85916)

justinchuby committed 2 years ago

9009393f

[MPS] Add support for aten::masked_select on mps (#119) (#85818)

DenisVieriu97 committed 2 years ago

6a14fcb9

Add mask_type=2 to masked_softmax for when mask.size() == input.size() (#85915)

george-qi committed 2 years ago

85258ec1

Fix rand_like nvprim meta function. (#85882)

kevinstephano committed 2 years ago

6004c65a

Update _torch_docs.py (#85924)

vfdev-5 committed 2 years ago

103a21f4

assert lambda >= 0 in poisson distribution cuda kernel (#85906)

yuguo68 committed 2 years ago

c036fb3e

Fix typo under docs directory and RELEASE.md (#85896)

kiszk committed 2 years ago

bc57306b

assert weights being 1-d tensor in bincount (#85881)

yuguo68 committed 2 years ago

11224f34

Revert "Improve make_tensor performance for float and complex types (#85473)"

pytorchmergebot committed 2 years ago

6db3539e

Align functorch docs with PyTorch's (#85856)

zou3519 committed 2 years ago

50000f3c

[functorch] Update notebooks for latest release (#85855)

zou3519 committed 2 years ago

19c7a6b5

Older