strisunshinewentingwang/pytorch

Pull Requests Commits

graph mode: add handling for layer_norm op (#37525)

vkuzo committed 4 years ago

a961d3ac

graph mode: add hardswish op (#37524)

vkuzo committed 4 years ago

4e7403c2

graph mode: refactor quantized hardswish API for easier graph handling (#37523)

vkuzo committed 4 years ago

7ac98c93

graph mode: add hardsigmoid op (#37522)

vkuzo committed 4 years ago

11b6f70f

graph mode: add elu op (#37521)

vkuzo committed 4 years ago

6cdc8cac

graph mode: add hardtanh op (#37469)

vkuzo committed 4 years ago

400098d4

[quant] Enable qnnpack tests for test_quantize and test_numeric_suite (#37351)

supriyar committed 4 years ago

b33b46a9

Cleanup internal functions in python_functions.cpp (#37536)

Shen Li committed 4 years ago

b48239af

Minor format cleanup in py_rref.cpp (#37520)

Shen Li committed 4 years ago

322e564e

Let RPC return FutureIValue instead of FutureMessage (#37519)

mrshenli committed 4 years ago

d5b38984

Enable float requantization for avgpool/gavgpool ops. (#37037)

kimishpatel committed 4 years ago

e9db51f9

Set onnx opset version before model select (#37466)

suffian khan committed 4 years ago

d5363e64

Make c10::complex the C++ type for complex tensors (#37421)

zasdfgbnm committed 4 years ago

1ef99263

Update assertion in MHA forward to support FP16 training (#37539)

cndn committed 4 years ago

5bb93573

Revert D21297549: [jit] fix trace checking reporting divergent names

suo committed 4 years ago

896f8130

`IValue(bool)` constructor should initialize entire payload (#37513)

malfet committed 4 years ago

f1cd0eeb

Migrate `cos` and `cos_` from TH to ATen (CUDA) (#36653)

kshitij12345 committed 4 years ago

7e9cc4df

Add `sched_setaffinity` check from libgomp to `valgrind.sup` (#37532)

malfet committed 4 years ago

6098cf7e

add support for generating Vandermonde matrices (#36725)

jessebrizzi committed 4 years ago

bca82801

Revert D21302691: [pytorch][PR] Implement cusparse Descriptor class and clean up cusparse code

ezyang committed 4 years ago

f7dce850

[quant] Enable convolution tests (#37494)

z-a-f committed 4 years ago

297cc551

Add overload names to dict operators. (#37279)

iseeyuan committed 4 years ago

ec5fb29b

Fix mobile type resolution in unpickling (#37425)

suo committed 4 years ago

9e97e924

Port xnnpack operators to new registration API (#36800)

ezyang committed 4 years ago

a3ab560f

Fix multiple issues with type annotations (#36358)

Bartosz Gasiorzewski committed 4 years ago

867e0592

Implement cusparse Descriptor class and clean up cusparse code (#37389)

xwang233 committed 4 years ago

bbf29a52

Extend some of the basic ops to kHalf (#37121)

malfet committed 4 years ago

1bb66a0c

Disable tests failing on test2 in ROCm CI (#37427)

ashishfarmer committed 4 years ago

bbd2350c

[cmake] add USE_SYSTEM_{XNNPACK,ONNX} options. (#37501)

cdluminate committed 4 years ago

58a46a17

Remove THCudaMemGetInfo. Use c10's cacheInfo instead. (#37447)

xuhdev committed 4 years ago

0d9e3b48

Newer Older