strisunshinewentingwang/pytorch

Pull Requests Commits

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

Inline coverage report combining/reporting (#49615)

malfet committed 3 years ago

6f381de0

[Issue #46210] added torch.fx.len() to provide support for len(); added a test case for torch.fx.len() (#49532)

huiguoo committed 3 years ago

e2e44bb1

[NNC] Disable masked fill (#49622)

Elias Ellison committed 3 years ago

3659560f

`torch.reciprocal`: promote integer inputs to float (#49102)

soulitzer committed 3 years ago

5ab95930

Output stacks (support for SVG visualization) (#48438)

Ilia Cherniavskii committed 3 years ago

485aee7a

Add sinc operator (#48740)

soulitzer committed 3 years ago

d0a12c5a

[torchscript] Fix constant propagation schemas (#49605)

IvanKobzarev committed 3 years ago

d088359e

Cleanup APIs for pipeline parallelism. (#48630)

pritamdamania committed 3 years ago

9d91360b

Upload test times to S3 (#49190)

samestep committed 3 years ago

39d89e06

[package] implicitly extern stdlib before mocking (#49306)

zdevito committed 3 years ago

b361e33a

[FX] Emit named tuple construction node when NamedTuple appears as an arg (#49553)

James Reed committed 3 years ago

fb755ad3

Test pipeline parallelism works with DDP. (#48470)

pritamdamania committed 3 years ago

27f355f8

Adding support for bitwise augassignment operators (#44621)

nikithamalgifb committed 3 years ago

e17f0fd6

New profiler API (#48280)

Ilia Cherniavskii committed 3 years ago

daaf932a

[PyTorch Mobile] Export Operator List from Mobile CompilationUnit instead of from TorchScript Model (#49385)

dhruvbird committed 3 years ago

4a870f65

Renaming CAFFE2_API to TORCH_API (#49496)

janeyx99 committed 3 years ago

71ca600a

[FX] Enforce args is tuple and kwargs is dict (#49526)

James Reed committed 3 years ago

c9e05213

Remove deadlines for Caffe2 hypothesis_test when running on GPU. (#49591)

Taylor Robie committed 3 years ago

faf60329

Fix Module backward hooks for all Tensor inputs/outputs (#46163)

albanD committed 3 years ago

ccd64669

fixed the first line of torch.rst to match the __init__.py file's first line (#49584)

jonykarki committed 3 years ago

0b27d570

Refactor VmapPhysicalView::newLogicalToPhysical (#49482)

zou3519 committed 3 years ago

7545ff66

add checkout PR tip step for quick checks (#49590)

Rong Rong (AI Infra) committed 3 years ago

f975f99d

Add op bench for caffe2 quantile op (#49598)

ShijunK committed 3 years ago

2de345d4

Newer Older