strisunshinewentingwang/pytorch

Pull Requests Commits

[qunat][fx][bc-breaking] Remove "additional_quant_pattern" key from prepare_custom_config_dict (#75386)

jerryzh168 committed 2 years ago

f42bdff0

[quant][fx] Using native backend_config_dict in fusion (#75378)

jerryzh168 committed 2 years ago

f26891c8

[quant][fx] Remove "additional_fusion_pattern" from prepare_custom_config_dict (#75377)

jerryzh168 committed 2 years ago

689cec94

[quant][fx] Move all fusion registrations to backend_config_dict (#75318)

jerryzh168 committed 2 years ago

dd667b6e

[shard] add ShardedTensor.cpu() (#74941)

wanchaol committed 2 years ago

0f7e60d6

Optimize size(dim) and stride(dim)

peterbell10 committed 2 years ago

e6842c0e

[fx2trt] improve to_dtype (#48)

Wei Wei committed 2 years ago

03dd22a2

[c10d] fix nccl gather outputs on non-root ranks (#75535)

wanchaol committed 2 years ago

caa11774

Disable strides and contiguity for CSR tensors

IvanYashchuk committed 2 years ago

3f1351d1

Revert D35489740: [pytorch][PR] Updated cudnn_frontend submodule to v0.6

David Dang committed 2 years ago

771527fc

[ci] Fix bug in get_workflow_job_id.py

suo committed 2 years ago

1fc2f4cc

[BE] Fix deprecated usages of `isIntegral`

malfet committed 2 years ago

6ba29d71

Add maximize flag to Adadelta

francescocastelli committed 2 years ago

58a44523

Docs: Fix `log_target` example in kl divergence

jaketae committed 2 years ago

c02bdfae

updated _forward_unim. to include descriptive error

zikpefu committed 2 years ago

43cc726c

Support masked sum on CSR tensors [CPU, CUDA]

pearu committed 2 years ago

e61b2e12

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

Newer Older