strisunshinewentingwang/pytorch

Pull Requests Commits

Add a NestedTensor Readme (#91472)

drisspg committed 1 year ago

eb8547e9

[Inductor] Support loop split at given depth in CPP codegen (#91397)

jgong5 committed 1 year ago

859ac58c

[inductor] fix output_stride of cat (#91233)

chunyuan-w committed 1 year ago

2555971b

[inductor] decompose tanh in CPP backend (#91687)

chunyuan-w committed 1 year ago

c99a2a43

Revert "[functorch] test: try using reference_inputs in vmap tests (#91355)"

pytorchmergebot committed 1 year ago

ad70a701

[functorch] test: try using reference_inputs in vmap tests (#91355)

kshitij12345 committed 1 year ago

a51090d4

Don't remove files across the whole OS on clean (#91503)

aostrowski-hbn committed 1 year ago

d0a4e2e7

[DTensor] fix test_device_mesh failure on GPU (#91783)

XilunWu committed 1 year ago

e3bd38d2

[ONNX] Support constant 'aten::__contains__' (#91660)

BowenBao committed 1 year ago

66745831

[MPS] Fix correctness issues with Pooling ops (#91519)

razarmehr committed 1 year ago

2f0e4839

inductor: Move graph.lint() in Intel's FX Passes to the End of Loop to Reduce Compile Time(part 2) (#91677)

XiaobingSuper committed 1 year ago

33547bb5

inductor:enable conv+unary fusion for torch unary function (#91609)

XiaobingSuper committed 1 year ago

25ff10ca

[cpu] implement erf based on oneDNN algorithm for aten::Vec (#91613)

Valentine233 committed 1 year ago

2175c941

[inductor] optimize lowering for empty-related operators (#91350)

Valentine233 committed 1 year ago

745dc3a1

Fix `test_math_ops` for python-3.11 (#91774)

malfet committed 1 year ago

e1a2b0d3

[Meta] Register aten.pixel_shuffle.default for meta (#91605)

Xia-Weiwen committed 1 year ago

de9c82f4

[Quant] Update IDeep to support oneDNN conv add fusion (#90605)

leslie-fang-intel committed 1 year ago

b2c68c1d

[Quant] Remove all the dequant nodes when the ref module has multi input args (#90157)

leslie-fang-intel committed 1 year ago

aab55d6d

Update version numbers in torch.{stft,istft} deprecations (#91761)

peterbell10 committed 1 year ago

ae0c4c4c

Fix rendering of std/var docs (#91730)

peterbell10 committed 1 year ago

2a64365a

Revert "Make torch.device usable as a context manager (#91525)"

pytorchmergebot committed 1 year ago

f571ae4f

Revert "[decomp] Use new squeeze.dims overload in decompositions (#91602)"

pytorchmergebot committed 1 year ago

c73147f7

feat: adding greater_equal Scalar variant (#91324)

srossross committed 1 year ago

0100293a

Make `requirements-ci.txt` reading cwd independent (#91771)

malfet committed 1 year ago

3b4e4d2b

training support for dynamo+torchxla integration (#88449)

shunting314 committed 1 year ago

a5f32f89

Revert "squeeze: allow squeezing multiple dimensions at once (#89017)"

pytorchmergebot committed 1 year ago

df4b3b13

[ROCm] tools/stats/monitor.py support (#91732)

jeffdaily committed 1 year ago

f11dc26e

[decomp] Use new squeeze.dims overload in decompositions (#91602)

peterbell10 committed 1 year ago

9262ffc6

Revert "Symintify pytorch slicing logic (#91340)"

pytorchmergebot committed 1 year ago

3bb63aa3

[MPS] Improve the performance of torch.linear() (#91114)

razarmehr committed 1 year ago

9ca37d65

Newer Older