strisunshinewentingwang/pytorch

Pull Requests Commits

Update onnx.rst

Nuno-Mota committed 2 years ago

0bd33545

[LT] Support diagonal op (#75230)

Jiewen Tan committed 2 years ago

dc37090e

Ensure convolution_backward respects output_mask

soulitzer committed 2 years ago

b10d1517

[PyTorch] NestedTensor kernels for {r,g}elu{,_}

swolchok committed 2 years ago

70cab5eb

[PyTorch] Add & use in-place gelu

swolchok committed 2 years ago

5c5a1e0a

[PyTorch] _addm_activation native function for matmul/bias/activation fusion

swolchok committed 2 years ago

48147675

Fix exporting models to ONNX without allow_tf32 in _convolution call

Eugene Lyapustin committed 2 years ago

d88a1160

Ansley's rebase of DimensionNode onto master (#75352)

Krovatkin committed 2 years ago

4a85145b

[complex32] enable testing for {h,v,d}stack

kshitij12345 committed 2 years ago

dfa1d953

Make split_with_sizes an overload of split

peterbell10 committed 2 years ago

8c1004e1

[complex] conv2d

kshitij12345 committed 2 years ago

fe799374

improve datapipe deprecation warnings (#74685)

pmeier committed 2 years ago

ff8705b3

Add edge case tests and update tril code to cover

drisspg committed 2 years ago

1d6007fa

ROCm: Enable test_masked_scatter_large_tensor

jaglinux committed 2 years ago

26ba7a92

Enable atomicAddNoRet() for all gfx targets.

hubertlu-tw committed 2 years ago

0d4d193a

fix 'pytorch/tools/code_coverage/README.md' for renamed options

sehoon committed 2 years ago

7a915a15

remove from generate_ci_workflows.py, add tags back in .github/templates

clee2000 committed 2 years ago

1bb6b3bd

[Quant][fx] Decouple prepare_*fx from training/eval modes (#75401)

andrewor14 committed 2 years ago

0bdf9a98

[quant][fx] Move embedding ops to backend_config_dict (#75317)

jerryzh168 committed 2 years ago

e4817e6c

forward-mode AD formula for F.dropout

zou3519 committed 2 years ago

68b18666

[ONNX] Raise exception for mixed precision input for BatchNormalization

Thiago Crepaldi committed 2 years ago

cbabd8f9

Updated cudnn_frontend submodule to v0.6 (#75481)

dzdang committed 2 years ago

6e11435c

move setup_helpers programs to their own package (#74838)

mikey dagitses committed 2 years ago

d92213f7

[quant][fx] Move rnn ops to backend_config_dict (#75316)

jerryzh168 committed 2 years ago

9905b1f2

[mobile] Update test model generation script to count op occurrences

linbinyu committed 2 years ago

c755d220

Add forward AD for rsub, polar, and FFT

IvanYashchuk committed 2 years ago

38a758e2

[quant] add checking number of args when checking observer in same graph (#75460)

terrychenism committed 2 years ago

37dea045

[wip] Correctly use pip3 and python3 for upload-test-stats

suo committed 2 years ago

8e95a2cd

Dynamo+LTC: handle inplace ops (#75359)

shunting314 committed 2 years ago

f9407fdb

[torch deploy] Add -rdynamic option explicitly to CMakeLists.txt (#75461)

s4ayub committed 2 years ago

9c56cc47

Newer Older