strisunshinewentingwang/pytorch

Pull Requests Commits

Implement torch.linalg.svd (#45562)

antocuni committed 3 years ago

5c5abd59

Update autograd related comments (#50166)

albanD committed 3 years ago

006cfebf

[numpy] torch.exp: promote integer inputs to float (#50093)

kshitij12345 committed 3 years ago

9f832c8d

Autograd engine, only enqueue task when it is fully initialized (#50164)

albanD committed 3 years ago

fc2ead09

Revert D25687465: [PyTorch] Devirtualize TensorImpl::dim() with macro

luciang committed 3 years ago

c215ffb6

Address clang-tidy warnings in ProcessGroupNCCL (#50131)

rohan-varma committed 3 years ago

294b7867

Disable cuDNN persistent RNN on sm_86 devices (#49534)

xwang233 committed 3 years ago

5a63c452

[PyTorch] Change representation of SizesAndStrides (#47508)

swolchok committed 3 years ago

b73c0185

[PyTorch] Introduce packed SizesAndStrides abstraction (#47507)

swolchok committed 3 years ago

882ddb2f

Completely remove FutureMessage type (#50029)

mrshenli committed 3 years ago

c480eebf

Completely Remove FutureMessage from RPC agents (#50028)

mrshenli committed 3 years ago

171648ed

Completely Remove FutureMessage from RPC cpp tests (#50027)

mrshenli committed 3 years ago

09875101

Remove FutureMessage from RPC request callback logic (#50026)

mrshenli committed 3 years ago

1f795e1a

Completely remove FutureMessage from FaultyProcessGroupAgent (#50025)

mrshenli committed 3 years ago

2831af98

Remove FutureMessage from sender TensorPipeAgent (#50024)

mrshenli committed 3 years ago

0684d074

Remove FutureMessage from sender ProcessGroupAgent (#50023)

mrshenli committed 3 years ago

1deb8950

Completely remove FutureMessage from distributed autograd (#50020)

mrshenli committed 3 years ago

0c943931

Completely remove FutureMessage from RPC TorchScript implementations (#50005)

mrshenli committed 3 years ago

b2da0b5a

Completely remove FutureMessage from RRef Implementations (#50004)

mrshenli committed 3 years ago

2d5f57cf

Replace FutureMessage with ivalue::Future in RpcAgent retry logic (#49995)

mrshenli committed 3 years ago

d730c7e2

Replace FutureMessage with ivalue::Future in RRefContext (#49960)

mrshenli committed 3 years ago

008206de

Replace FutureMessage with ivalue::Future in distributed/autograd/utils.* (#49927)

mrshenli committed 3 years ago

25ef6051

Let RpcAgent::send() return JitFuture (#49906)

mrshenli committed 3 years ago

84e3237a

[PyTorch] Devirtualize TensorImpl::dim() with macro (#49770)

swolchok committed 3 years ago

4de6b279

[PyTorch] validate that SparseTensorImpl::dim needn't be overridden (#49767)

swolchok committed 3 years ago

1a1b6658

[PyTorch] Devirtualize TensorImpl::numel() with macro (#49766)

swolchok committed 3 years ago

2e7c6cc9

Fix SyncBatchNorm usage without stats tracking (#50126)

malfet committed 3 years ago

bf4fcab6

add type annotations to torch._utils (#49705)

guilhermeleobas committed 3 years ago

870ab04b

[Gradient Compression] Remove the extra comma after "bucket" in PowerSGD hook signatures (#50197)

Yi Wang committed 3 years ago

ce370398

Clean up some type annotations in android (#49944)

r-barnes committed 3 years ago

09eefec6

Newer Older