strisunshinewentingwang/pytorch

Pull Requests Commits

fix _unsafe_view schema to work with functionalization

bdhirsh committed 2 years ago

46234df5

support horizontal fusion

SherlockNoMad committed 2 years ago

04d26947

[JIT] Propagate profiled information to DifferentiableGraph outputs

davidberard98 committed 2 years ago

1d2a6c2e

Support multi-dimensional lengths in segment_reduce to support pytorch_scatter.segment_* functionalities (CUDA)

mikaylagawarecki committed 2 years ago

40f7ef1f

Expose _export_data from C++ to Python (#79207)

Luka Mushkudiani committed 2 years ago

c0a7c1d0

test_fuser_util passing

SherlockNoMad committed 2 years ago

8cc6102a

Revert "[ci] remove IS_GHA env var"

pytorchmergebot committed 2 years ago

338bfe63

Delete syncbranches script/workflow (#79245)

malfet committed 2 years ago

01929b73

[ci] remove IS_GHA env var

suo committed 2 years ago

1a2d95c6

[ci] remove COMPACT_JOB_NAME env var

suo committed 2 years ago

1e2890ab

[ci] remove ppc64ie build env

suo committed 2 years ago

0a4d6465

[ci] fix USE_DEPLOY condition

suo committed 2 years ago

18c3314e

Fix how we handle host memory in CUDA getDeviceFromPtr

cbalioglu committed 2 years ago

21be4d40

multiple output working

SherlockNoMad committed 2 years ago

9e2a0ab4

[CI] Turn flaky test signal to green (#79220)

janeyx99 committed 2 years ago

1bc8c873

Add Dynamic Output Shape Tagdfor ata-dependent ops, handle in FakeTensor

eellison committed 2 years ago

13a8867c

Remove Unused Nervana GPU submodule (#79168)

osalpekar committed 2 years ago

f221b16b

FX fuse by partition

SherlockNoMad committed 2 years ago

708839fe

Add onnx / test to required merge rules (#78790)

justinchuby committed 2 years ago

54ea6bbc

Revert "[mergebot] Make merge on green default behavior (#79199)"

pytorchmergebot committed 2 years ago

39bfdf41

[mergebot] Make merge on green default behavior (#79199)

zengk95 committed 2 years ago

ddeeca0b

Support both train / eval modes for ModuleInfo

jbschlosser committed 2 years ago

70d6446a

Minor FX test fix for TorchDynamo (#79206)

anijain2305 committed 2 years ago

79f18c1a

Properly setup __name__ on refs functions.

ezyang committed 2 years ago

b18ba7e0

Move test_profiler tests to tree rather than icicle format

Taylor Robie committed 2 years ago

84b9e5ba

Revert "Revert "[Profiler] Move python tracing to unified event type (Part 2)""

Taylor Robie committed 2 years ago

9f2e2aa2

Add onnx support for `movedim` and `moveaxis` (#78931)

qqaatw committed 2 years ago

2bafb42a

Fix jit schema_matching ignoring self resulting in wrong operator schema

yuguo68 committed 2 years ago

c1b831f9

Support multi-dimensional lengths in segment_reduce to support pytorch_scatter.segment_* functionalities (CPU-only)

mikaylagawarecki committed 2 years ago

e289a18e

[quant][core][feature] Implemented index_put for QuantizedCPU tensors

dzdang committed 2 years ago

e40e3781

Newer Older