strisunshinewentingwang/pytorch

Pull Requests Commits

Sparse-sparse matrix multiplication (CPU/CUDA) (#39526)

aocsa committed 3 years ago

44ce0b88

Implementing NumPy-like function torch.broadcast_to (#48997)

RockingJavaBean committed 3 years ago

3779bdec

[NNC] Intermediate allocs flattened and dependency support (#49554)

nickgg committed 3 years ago

db2e9c1e

Fixed a typo in dataloader.py. (#49437)

tmcclintock committed 3 years ago

a3aafea0

Fix typo in add_pr_curve docstrings. (#49648)

theodumont committed 3 years ago

b1a1271f

Fix return type Any for Ternary ops (#49165)

ejguan committed 3 years ago

b80a3661

Added linalg.solve (#48456)

IvanYashchuk committed 3 years ago

8be205ae

Fix sinc docs typo (#49667)

soulitzer committed 3 years ago

5ce94991

[Mask R-CNN]Add Int8 AABB Generate proposals Op (#49574)

anshuljain1 committed 3 years ago

ef172e13

[WIP][DataLoader] Prototype of SamplerIterableDataset (#49363)

ejguan committed 3 years ago

7ed140a1

[WIP][DataLoader] Prototype of BatchIterableDataset (#49186)

ejguan committed 3 years ago

554f79ac

[WIP][DataLoader] CollateIterableDataset prototype (#48933)

ejguan committed 3 years ago

1b6fc1fd

[AutoAccept][Codemod][FBSourceClangFormatLinter] Daily `arc lint --take CLANGFORMAT`

generatedunixname89002005325676 committed 3 years ago

bab732a3

Add support for torch.tensor_split to accept a tensor for `indices` argument (#49169)

Edson Romero committed 3 years ago

5c3788d5

[Gradient Compression] Replace the assertions in PowerSGD comm hook by stream syncrhonization (#49435)

Yi Wang committed 3 years ago

96aed203

[Gradient Compression] Add error feedback to layerwise PowerSGD (#49418)

Yi Wang committed 3 years ago

342bfd89

stft: Change require_complex warning to an error (#49022)

peterbell10 committed 3 years ago

5c25f8fa

Updated derivative rules for complex svd and pinverse (#47761)

IvanYashchuk committed 3 years ago

f5ee619d

Resubmit: [Gradient Compression] Implement the original layerwise PowerSGD (#49639)

Yi Wang committed 3 years ago

8b61fbda

disable kthvalue overlap (#48254)

guol-fnst committed 3 years ago

c0deb231

Remove DataPtr extractor from CUDAFuture (#48840)

lw committed 3 years ago

1ac05cfe

Disable test on windows (#49636)

Ilia Cherniavskii committed 3 years ago

e0f60c97

[PyTorch Mobile] Preserve bundled input related methods when calling optimize_for_mobile (#49170)

bearzx committed 3 years ago

e2d2d9bb

Revert D25511543: [Gradient Compression] Implement the original layerwise PowerSGD

mrshenli committed 3 years ago

ad9923e5

[quant][graphmode][fx] Allow user to specify qconfig for call_method (#49621)

jerryzh168 committed 3 years ago

5cde23fd

Fix lint (#49629)

mrshenli committed 3 years ago

e4eaa6de

Bump tensorpipe version (#49599)

Lucas Hosseini committed 3 years ago

7278e3bd

Add benchmark for torch.distributed.pipeline.sync.Pipe (#49577)

pritamdamania committed 3 years ago

159de1f1

Improve documentation for pipeline parallelism. (#48638)

pritamdamania committed 3 years ago

8c52fdf5

[Gradient Compression] Implement the original layerwise PowerSGD (#49417)

Yi Wang committed 3 years ago

71f3399e

Newer Older