strisunshinewentingwang/pytorch

Pull Requests Commits

Adjust computeLRWorkDim() to work with Accelerate.framework (#56847)

malfet committed 3 years ago

ed617a61

Add dispatch keys for out-of-tree grad+vmap prototype (#56824)

zou3519 committed 3 years ago

338a600e

add all pools, Batchnorm and Tanh (i.e. all ideeped MKLDNN ops) to MKLDNNFuser (#56541)

Krovatkin committed 3 years ago

cfbd06d7

.github: Bump linux.2xlarge runners to 500 (#56945)

seemethere committed 3 years ago

8d29ac20

.github: Build test binaries in build/ directory (#56941)

seemethere committed 3 years ago

e1389878

[NNC] removed the second run of llvm passmanager - it is repeated and caused a slowdown in the generated code (#56837)

huiguoo committed 3 years ago

6bbd8ba6

[DataLoader] Add `generate_state` for NumPy seeding (#56797)

ejguan committed 3 years ago

3b977a0d

add missing comma to `run_test.py` (#57010)

pmeier committed 3 years ago

759cfb74

Enable fixed fast_mode for complex (#55699)

soulitzer committed 3 years ago

201ad938

Refactor C->C to C->R twice (#55692)

soulitzer committed 3 years ago

7fe6e8e5

Add OpInfos for torch.{complex, view_as_real, view_as_complex} (#56524)

anjali411 committed 3 years ago

268cc117

Added OpInfo for torch.einsum (#56276)

heitorschueroff committed 3 years ago

57e37080

Remove C++17 only optional include (#56782)

ezyang committed 3 years ago

ab1457ad

Make test_randperm work with meta device (#56976)

ezyang committed 3 years ago

0d777a80

Implement module.to_empty() (#56610)

jbschlosser committed 3 years ago

f7fba854

DOC: Add note to mutating methods (#56877)

bharatr21 committed 3 years ago

f2acdff7

Revert D27831996: ns for fx: move node I/O dtype mapping to be local instead of global

Mike Ruberry committed 3 years ago

1145e2c6

Revert D27833189: ns for fx: allow user functions in shadowing

Mike Ruberry committed 3 years ago

45e96b54

Revert D27836064: ns for fx: add fp16 function shadowing

Mike Ruberry committed 3 years ago

982c72ac

Revert D27857735: ns for fx: bug fix for shadowing fp16 emulation patterns

Mike Ruberry committed 3 years ago

90d554bd

Revert D27864296: ns for fx: support binary ops when adding unshadowed loggers for inputs

Mike Ruberry committed 3 years ago

abb8b6c1

Revert D27886107: ns for fx: add option to skip matching classes and functions

Mike Ruberry committed 3 years ago

cc8c5c14

Revert D27960767: ns for fx: allow comparing int8 to int8 for functionals

Mike Ruberry committed 3 years ago

5dc7a6b0

Revert D27960766: ns for fx: additional bugfix for user defined functions

Mike Ruberry committed 3 years ago

5db03b41

Back out "fx: Fix type_matches for Optional[List[int]] arguments" (#56991)

spaugh committed 3 years ago

a0483cd0

Add some functions for manipulating mkldnn tensors to TORCH_API (#56954)

bertmaher committed 3 years ago

780f4542

Revert "Use at::cpu in bench_approx (#56563)" (#56816)

bertmaher committed 3 years ago

c42dd8b2

[profiler] Add cuda synchronization points (#56651)

Ilia Cherniavskii committed 3 years ago

38bb0ac3

Move caffe2 signal_handler to c10. (#56717)

pritamdamania committed 3 years ago

dc8a8cea

[TensorPipe] Give higher priority to CPU-only channels. (#56908)

Lucas Hosseini committed 3 years ago

6ed5bbfb

Newer Older