strisunshinewentingwang/pytorch

Pull Requests Commits

[ONNX] Fix exporting copy_ with index as tensor input (#32801)

BowenBao committed 4 years ago

432858c9

[JIT] Add Exit Transform / Convert To SSA to docs

Elias Ellison committed 4 years ago

ca33aeba

avinashsai committed 4 years ago

b0476dc6

[JIT] Resolve custom classes in source importer (#32977)

James Reed committed 4 years ago

38820a70

[c10] Allow taking a std::tuple as arg (#32948)

James Reed committed 4 years ago

757cea92

Revert D19730209: [pytorch][PR] Issue a warning when using zero_grad in DataParallel

zou3519 committed 4 years ago

8195961f

Revert D19417087: fix #30480 torch.normal shape checking is broken

zou3519 committed 4 years ago

ec1e9a1a

[C2] Introduce extra_info force CPU tags for auto-generated iteration counter blobs (#32607)

kennyhorror committed 4 years ago

e76fa982

fix #30480 torch.normal shape checking is broken (#32243)

lixinyu committed 4 years ago

3c17cbb6

Move normal distribution to Aten(CPU) (#32031)

XiaobingSuper committed 4 years ago

b00345a6

Issue a warning when using zero_grad in DataParallel (#32870)

peterbell10 committed 4 years ago

46c3c18b

Add option to use ninja to compile ahead-of-time cpp_extensions (#32495)

zou3519 committed 4 years ago

62094126

[ONNX] Add flag to enable script tests (#32654)

BowenBao committed 4 years ago

e54d9545

Consider hub_dir alongside TORCH_HOME env variable for storing hub models (#32844)

andfoy committed 4 years ago

1b746b95

Fix some bugs with zipfile serialization (#32244)

davidriazati committed 4 years ago

74ce3a03

Add ability to abort NCCL communicators from the store. (#32895)

pritamdamania committed 4 years ago

ab75d64e

[jit] fix parser for one-line functions (#32941)

suo committed 4 years ago

df1d68d5

Enabling the nccl/rccl test for ROCM environment (#32340)

mpruthvikumar committed 4 years ago

908b451e

Properly update _flat_weights in RNN models (#32989)

ngimel committed 4 years ago

e8581869

Move some broadcasting logic away from codegen. (#32982)

gchanan committed 4 years ago

72b9412b

[aten] fix vector memory leak (#32478)

gaurav1086 committed 4 years ago

fbde3c05

Fix dispatch of argmax/argmin. (#32961)

Ailing Zhang committed 4 years ago

81a90463

Kill _th_max, _th_min overloads that aren't used.

gchanan committed 4 years ago

3531f993

Add XLAPreAutograd key for XLA use cases that need custom autograd. (#32788)

ezyang committed 4 years ago

16c166e2

Stop using dispatchTypeId to do checks for tensor list unwrap. (#32787)

ezyang committed 4 years ago

6b0813ea

Expose Channel Last 3d enum

lixinyu committed 4 years ago

1b446aa2

Attempt to workaround MSVC17 static constexpr bug

James Reed committed 4 years ago

836b4c9e

[JIT] Fix python pickle serialization for torchbind (#32878)

James Reed committed 4 years ago

f393adc0

[JIT] Make IRParser use op schema (#32854)

James Reed committed 4 years ago

23a48007

[JIT] Trace uses of torchbind classes as module attributes (#32833)

James Reed committed 4 years ago

bc4790b3

Newer Older