strisunshinewentingwang/pytorch

Pull Requests Commits

[NNAPI] Fix models with no weights (#47517)

dreiss committed 3 years ago

1be909f0

Make bundled inputs work with quantized zero inputs (#47407)

dreiss committed 3 years ago

0e7af36a

[vulkan] Add Winograd convolutions (#54639)

SS-JIA committed 3 years ago

ad5dc84e

[Pytorch Mobile] Fold Conv BatchNorm for functions besides forward (#54619)

JacobSzwejbka committed 3 years ago

20d7916a

Revert back changes in test_custom_ops.cpp. (#55350)

Ailing Zhang committed 3 years ago

a9bcab46

[JIT] Include conv3d in the conv-add-relu fusion (#54772)

desertfire committed 3 years ago

a0d97761

Revert D27246997: [pytorch][PR] Fix reference cycle in sparse coalesce graph

bdhirsh committed 3 years ago

ec80981d

Revert D27572158: [torchelastic] Make sure torchelastic mp wait for queue to be drained before finishing the process

bdhirsh committed 3 years ago

ae3a876c

[Resubmit] Fix for incorrect usage of logging in torch/distributed/distributed_c10d.py (#52757)

szmigacz committed 3 years ago

8e78a1b0

[torchelastic] Make sure torchelastic mp wait for queue to be drained before finishing the process

aivanou committed 3 years ago

e9c6a511

Upgrade onednn to v2.1.2 (#54956)

XiaobingSuper committed 3 years ago

f8788d51

Add MonkeyType dependency for testing on Linux (#55305)

nikithamalgi committed 3 years ago

8243ba72

Correct many OpInfos "test_out" skips. (#55141)

lezcano committed 3 years ago

158cdece

Fix reference cycle in sparse coalesce graph (#52874)

peterbell10 committed 3 years ago

815bfad2

Optimized generic interpolation using TensorIterator (keeps original 2d/3d channels last impl) (#54500)

vfdev-5 committed 3 years ago

e5f66f00

Fix the clang-tidy diff SHA for using PR merge (#55318)

samestep committed 3 years ago

87d55058

Revert D27442325: [torch/elastic] Revise the rendezvous handler registry logic.

bdhirsh committed 3 years ago

bf70fe69

[Static Runtime] Make sure the copy version of the op exist in ReplaceWithCopy (#55337)

Hao Lu committed 3 years ago

c3d0607f

[Gradient Compression] Update _powerSGD_comm_hook_wrapper to only expose 2 most critical hyperparameters (#55295)

Yi Wang committed 3 years ago

1b4bb369

[Gradient Compression] Update the default value of start_powerSGD_iter and update the docstring (#55272)

Yi Wang committed 3 years ago

cc403690

[PyTorch] Lite interpreter with a backend delegate (#54462)

iseeyuan committed 3 years ago

3551bd31

[PyTorch] Fix bin hash comparison failure in clang format script (#55281)

cccclai committed 3 years ago

7d9a6197

[torch/elastic] Revise the rendezvous handler registry logic.

cbalioglu committed 3 years ago

df299dbd

[torch/elastic] Improve the implementation of `RendezvousParameters` and add its unit tests. (#146)

cbalioglu committed 3 years ago

359d0a02

[torch/elastic] Improve the implementation of the utility functions and add their unit tests. (#54804)

cbalioglu committed 3 years ago

7f06c65a

[torch/elastic] Expose a `stderr` parameter in `EtcdServer`. (#54805)

cbalioglu committed 3 years ago

de7f05b9

[torch/elastic] Revise the rendezvous exception types. (#54803)

cbalioglu committed 3 years ago

bad8d347

Wrap cub in its own namespace (#55292)

ngimel committed 3 years ago

55843321

[DDP] Call ensure_prior_reduction_finished within lock (#55074)

rohan-varma committed 3 years ago

0e03a297

Add some missing types to torch (#55184)

r-barnes committed 3 years ago

697b1303

Newer Older