strisunshinewentingwang/pytorch

Pull Requests Commits

Prevent sum overflow in broadcast_object_list (#70336)

Stephan Uphoff committed 2 years ago

807f9a82

Automated submodule update: tensorpipe (#70438)

facebook-github-bot committed 2 years ago

5a9ea9e3

Back out "Make TorchScript Preserve Fully Qualified Class Name for Python Exceptions"

Bo Wu committed 2 years ago

bf610f08

Add graph op replacement pass (#69915)

tugsbayasgalan committed 2 years ago

4ae71c8d

Extend Graph, CompilationUnit, and schema matching to accept optional operator version number (#69914)

tugsbayasgalan committed 2 years ago

63e58d26

Add utility methods to find an upgrader (#68355)

tugsbayasgalan committed 2 years ago

df3cbcff

Make TorchScript Preserve Fully Qualified Class Name for Python Exceptions (#70339)

shunting314 committed 2 years ago

911d527b

[Compiled Mobilenetv3 Demo] Integrate Compiled Mobilenetv3 into FB4A Playground app (#70370)

priyaramani committed 2 years ago

ab4f9862

Revert D33234529: [NNC Testing] Randomized loop nest infrastructure

Mikhail Zolotukhin committed 2 years ago

0ee663d2

Allow single node fusion for nvfuser (#70000)

jjsjann123 committed 2 years ago

e429a684

Do not use ZeroTensor for inplace ops (#69998)

soulitzer committed 2 years ago

5ccf28d0

Add forward AD formulas for `{adaptive_,fractional_,}max_pool{2,3}d_{backward,}` (#69884)

soulitzer committed 2 years ago

3116d870

[acc_ops] No longer mark acc_ops.cat as unary (#70365)

jfix71 committed 2 years ago

6925576e

Revert D33301254: [pytorch][PR] GHA Windows: Propagate exit code from .bat to calling bash script

janeyx99 committed 2 years ago

133c7f2c

GHA Windows: Propagate exit code from .bat to calling bash script (#70011)

janeyx99 committed 2 years ago

6431ac6c

[LTC] Upstream utils to extract BackendDevice from at::Tensor (#70069)

Jiewen Tan committed 2 years ago

ab57f6d1

[Easy] Lint wrap.py file (#70341)

rohan-varma committed 2 years ago

16e6e1a5

[FSDP] Remove module.wrapper_config support (#70340)

rohan-varma committed 2 years ago

3c231e9b

`torch.linalg` routines return `torch.linalg.LinAlgError` when a numerical error in the computation is found. (#68571)

v0dro committed 2 years ago

d100d98d

[SR] Fast path for VarStack on scalars (#70210)

Mike Iovine committed 2 years ago

6a844492

Transformer{DecoderLayer} : no batch dim (#70322)

kshitij12345 committed 2 years ago

cc8b9163

GaussianNLLLoss no_batch_dim docs and testing (#69783)

george-qi committed 2 years ago

4d49af86

Add the `maximize` flag to AdamW (#70146)

Adnios committed 2 years ago

a9c7d626

enable backward pass computation and communication overlap by prefetching all gather (#70235)

zhaojuanmao committed 2 years ago

b15212c6

[NNC Testing] Randomized loop nest infrastructure (#70174)

anijain2305 committed 2 years ago

1d094587

[quant][fx][graphmode] Add backend_config_dict for standalone module (#70150)

jerryzh168 committed 2 years ago

656d2a7b

Revert D33172665: [LTC] Upstream utils to extract BackendDevice from at::Tensor

suo committed 2 years ago

795af157

update poisson_nll_loss opinfo samples (#70300)

kshitij12345 committed 2 years ago

12afe2bb

[Profiler] Address issues from profiler bifurcation. (#70327)

Taylor Robie committed 2 years ago

681e78ba

[LTC] Upstream utils to extract BackendDevice from at::Tensor (#70069)

Jiewen Tan committed 2 years ago

121d0679

Newer Older