strisunshinewentingwang/pytorch

Pull Requests Commits

Add nondeterministic check to new upsample overloads

dreiss committed 4 years ago

df6ea625

Fix L1Loss when target.requires_grad is True. (#44471)

gchanan committed 4 years ago

3de2c0b4

[dper3] Export PackSegments and UnpackSegments to Pytorch

Brandon Lin committed 4 years ago

ea558206

[PyTorch Mobile] Move some string ops to register_prim_ops.cpp and make them selective (#44500)

iseeyuan committed 4 years ago

b73b44f9

In common_distributed, fix TEST_SKIPS multiprocessing manager (#44525)

rohan-varma committed 4 years ago

567c51cc

Fix MSELoss when target.requires_grad is True. (#44437)

gchanan committed 4 years ago

d07d25a8

Update submodule gloo to have latest commits to enable it can work on Windows (#44529)

gunandrose4u committed 4 years ago

9a3b83cb

torch.view_as_complex fails with segfault for a zero dimensional tensor (#44175)

guol-fnst committed 4 years ago

b6b1c01a

Use TP Tensor.metadata to carry device info (#44396)

mrshenli committed 4 years ago

a9754fb8

Add missing rpc.shutdown() (#44417)

mrshenli committed 4 years ago

f44de7cd

C++ APIs Transformer NN Module Top Layer (#44333)

lixinyu committed 4 years ago

77cc7d1e

Clarify track_running_stats docs; Make SyncBatchNorm track_running_stats behavior consistent (#44445)

ssnl committed 4 years ago

09892de8

[NNC] Don't attempt to refactor conditional scalars (#44223)

nickgg committed 4 years ago

30fccc53

[quant] conv_transpose1d_prepack / conv_transpose1d_unpack (#40360)

z-a-f committed 4 years ago

c967e772

[JIT] Remove profiling nodes in autodiff forward graph (#44420)

Elias Ellison committed 4 years ago

8b898666

[JIT] Add a python hook for a function to interpret JIT graphs. (#44493)

Mikhail Zolotukhin committed 4 years ago

c6febc64

Replace FutureMessage with c10::ivalue::Future in DistEngine. (#44239)

pritamdamania committed 4 years ago

51ed3126

Enable lerp on half type; fix output memory format (#43541)

xwang233 committed 4 years ago

b5d75ddd

[quant][eagermode][refactor] Add set/get method for quantization and fusion mappings (#43990)

jerryzh168 committed 4 years ago

0c58a017

[NCCL] Fix NCCL_BLOCKING_WAIT functionality with Async Error Handling (#44411)

osalpekar committed 4 years ago

f7278473

Speedup Linux nightly builds (#44532)

malfet committed 4 years ago

6ee41974

Register diag_backward, diagonal_backward, infinitetely...gelu_backward as operators (#44422)

zou3519 committed 4 years ago

69f6d94c

Register cummaxmin_backward, cumprod_backward as operators (#44410)

zou3519 committed 4 years ago

7ff7e6cf

Add trace_backward, masked_select_backward, and take_backward as ops (#44408)

zou3519 committed 4 years ago

08b431f5

Fix DDP join() API in the case of model.no_sync() (#44427)

rohan-varma committed 4 years ago

41f62b17

Fix uniqueness check in movedim (#44307)

peterbell10 committed 4 years ago

129d52ae

Moves some of TestTorchMathOps to OpInfos (#44277)

Mike Ruberry committed 4 years ago

c48f511c

Support work.result() to get result tensors for allreduce for Gloo, NCCL backends (#43970)

mrzzd committed 4 years ago

2e744b18

Disable PyTorch iOS ARM64 builds until cert problem is fixed (#44499)

malfet committed 4 years ago

91b16bff

[pytorch] Add logging to mobile Method run (#44234)

ann-ss committed 4 years ago

1dd3fae3

Newer Older