strisunshinewentingwang/pytorch

Pull Requests Commits

[Gradient Compression] Add is_the_last_bucket_to_allreduce method to GradBucket class (#53010)

Yi Wang committed 3 years ago

b05dd931

[Gradient Compression] Don't provide default values in GradBucket constructor (#53102)

Yi Wang committed 3 years ago

4997c38a

[Gradient Compression] Add get_per_parameter_tensors method to GradBucket class (#53009)

Yi Wang committed 3 years ago

ecb5ac90

Add default arguments to cuda stream and events (#53025)

nikithamalgifb committed 3 years ago

ab7f6f3f

Add wait_for_worker param to TCPStore and fix port in use flaky test failures (#52888)

H-Huang committed 3 years ago

2444b4d1

Store coverage files as artifacts for better debugging (#53126)

janeyx99 committed 3 years ago

41765d46

Add a note in DDP doc to point to ZeroRedundancyOptimizer (#53113)

mrshenli committed 3 years ago

d6970902

[Reland] Update and expose ZeroRedundancyOptimizer docs (#53112)

mrshenli committed 3 years ago

29034b94

[CUDA graphs] [JIT] improves readability and nvfuser convenience for graph-safe cuda RNG (#51580)

mcarilli committed 3 years ago

66b20bb7

Register DefaultBackend implementations for functional/inplace structured operators (#53037)

ezyang committed 3 years ago

37bf6c13

Fix minor inaccuracy in translate error reporting (#53032)

ezyang committed 3 years ago

c5a67f16

Revert D26733731: [pytorch][PR] Skip dispatch for `is_floating_point`

ezyang committed 3 years ago

fbf2883d

Clang-format quantization_hooks.py (#53100)

Yi Wang committed 3 years ago

890e0510

[operator_benchmark] Added channels last 3d option to interpolate test (#53117)

vfdev-5 committed 3 years ago

cb1596a1

Automated submodule update: tensorpipe (#53012)

facebook-github-bot committed 3 years ago

62d1cdd7

Revert D26753571: [pytorch][PR] add submodules to sys.modules so their attributes can be pickled

ezyang committed 3 years ago

2d7119f9

disable dill extension behavior (#53118)

suo committed 3 years ago

73a57246

Add streams boundary check to `torch::cuda::scatter`` (#53057)

malfet committed 3 years ago

43f810fa

fix logcumsumexp functor to properly handle infs and nans (#52947)

ngimel committed 3 years ago

e5e54ada

[ROCm] Enable test cases in test_nn.py for ROCm (#52836)

KyleCZH committed 3 years ago

d8ef3a47

Remove useless test_reference_numerics skip infos (#52890)

zasdfgbnm committed 3 years ago

2bf079d0

add submodules to sys.modules so their attributes can be pickled (#53107)

mattip committed 3 years ago

fbf9745c

add OpInfo entry for signbit (#52198)

kshitij12345 committed 3 years ago

aa603cb2

Skip dispatch for `is_floating_point` (#52998)

iramazanli committed 3 years ago

4fb82a88

[static runtime] Register both TupleConstruct and ListConstruct as out variants (#52684)

bwasti committed 3 years ago

d4e64dad

[static runtime] Add Alias analysis to Memory Management/Planning (#50060)

bwasti committed 3 years ago

2d67b76f

Explicitly export submodules and variables from torch module (#52339)

jakebailey committed 3 years ago

b22df263

Add duplicate scheduled-ci to allow for debugging (#53109)

janeyx99 committed 3 years ago

048e3917

Don't run cpp tests a second time in the sharded ort_test2 job (#53067)

janeyx99 committed 3 years ago

28f87bb7

Store test file in S3 as well for every TestSuite (#52869)

janeyx99 committed 3 years ago

09ce9b58

Newer Older