strisunshinewentingwang/pytorch

Pull Requests Commits

Update CODEOWNERS

albanD committed 3 years ago

Verified 0c54f118

FIX Validates target in cosine_embedding (#53110)

thomasjpfan committed 3 years ago

e2ecfb60

Revert D26720919: [Gradient Compression] Remove some low-level methods of GradBucket class

osalpekar committed 3 years ago

593b0fba

Suppress unsigned comparison warning (#52653)

r-barnes committed 3 years ago

c4c20a5d

Update torch.nn.quantizable.MultiHeadAttention docstring (#53106)

nihui committed 3 years ago

6ab3a8b6

Codegen python bindings to access attributes of grad_fn (#52451)

soulitzer committed 3 years ago

a3a21504

Back out "Revert D26753571: [pytorch][PR] add submodules to sys.modules so their attributes can be pickled" (#53127)

ezyang committed 3 years ago

baed2cfe

[Gradient Compression] Remove some low-level methods of GradBucket class (#53098)

Yi Wang committed 3 years ago

521e1e83

[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

Older