strisunshinewentingwang/pytorch

Pull Requests Commits

Skip copy_same_type_transpose_ for quantized tensor (#29609)

jerryzh168 committed 5 years ago

82913a26

Benchmarking per channel quantization

z-a-f committed 5 years ago

3b43cfde

Quantized interpolation benchmarks

z-a-f committed 5 years ago

5db361bd

Separate out pytorch_jni into pytorch_jni_jit and pytorch_jni_common. (#29617)

xcheng16 committed 5 years ago

9c9c361f

Benchmarking quantized methods (#29625)

z-a-f committed 5 years ago

f95e8ea1

Suppress implicit int-float conversion warning in ROCm build (#29604)

bddppq committed 5 years ago

f111f1b1

Fix jit tracing namedtuple (#29477)

bddppq committed 5 years ago

949d6ae1

Complex support on GPU for dynamic casting (#29612)

zasdfgbnm committed 5 years ago

450949c7

Enable test_distributed for ROCm but only with nccl backend

jithunnair-amd committed 5 years ago

7073ee20

quantized topk benchmarking

z-a-f committed 5 years ago

3b452ca4

Quantized unary ops benchmarking (mostly template)

z-a-f committed 5 years ago

a0d4d506

Updating submodules

svcscm committed 5 years ago

e651494d

change `drop_on_export` warning category (#29610)

suo committed 5 years ago

2fb40596

Convert conv_prepack to conv2d_prepack and conv_unpack to conv2d_unpack (#29529)

jianyuh committed 5 years ago

bbff06ee

Add Support for ONNX scripting Interpolate with missing shape (#29489)

lara-hdr committed 5 years ago

2acca09e

Updating submodules

svcscm committed 5 years ago

8db06732

Apply the latest master docker images(jni.h in every image) (#29588)

IvanKobzarev committed 5 years ago

0c9e6727

Add ONNX Export Support for torch.scalar_tensor (#28713)

lara-hdr committed 5 years ago

8b53515b

Disable android gradle jobs (#29606)

IvanKobzarev committed 5 years ago

5249c43d

Creating a base benchmarking class for activations.

z-a-f committed 5 years ago

fb07098e

explicitly provide memory format when calling to clone() at prune.py (#29593)

ifedan committed 5 years ago

7df854bd

explicitly provide memory format when calling to *_like operators

VitalyFedyunin committed 5 years ago

bf61405e

explicitly provide memory format when calling to *_like operators

VitalyFedyunin committed 5 years ago

8df60240

Updating submodules

svcscm committed 5 years ago

858b2010

Back out "Revert D18299298: [pytorch][PR] Migrate conv3d from TH to ATen (CPU)" (#29286)

ezyang committed 5 years ago

1bb5209f

Revoking mutually exclusive requirement on channels last and contiguous tensor (#28466)

VitalyFedyunin committed 5 years ago

ddeeb561

Revert D18253777: Remove observer module after insert_quant_dequant

jerryzh168 committed 5 years ago

70f886ff

change op bench input shape to reduce execution time (#29616)

mingzhe09088 committed 5 years ago

af3468a1

remove SkipInputShape flag (#29615)

mingzhe09088 committed 5 years ago

7374dd0d

Identify weights and bias by argument position in aten call (#29147)

jerryzh168 committed 5 years ago

fdcb203e

Newer Older