strisunshinewentingwang/pytorch

Pull Requests Commits

Move resize logic for bmm from codegen to native code. (#37955)

gchanan committed 4 years ago

b2cc9928

Acquire GIL when constructing/destructing ConcretePyObjectHolder (#37870)

mrshenli committed 4 years ago

ee1ddcef

Add support for non-persistent buffers. (#37191)

sharvil committed 4 years ago

594b33ea

Add --check-untyped-defs to mypy.ini and test suite (#37594)

rgommers committed 4 years ago

46ed3349

Migrate CUDA where, tril, triu to c10::complex (#37896)

zasdfgbnm committed 4 years ago

30fc58cf

[ONNX] Support clamp_min and clamp_max (#37872)

BowenBao committed 4 years ago

7be9796c

[TensorPipe] Use the new multi-payload message API (#37919)

lw committed 4 years ago

bc09478a

[TensorPipe] Allow passing args to agent options constructor (#37918)

lw committed 4 years ago

978ad162

remove deprecation warning on get_contiguous_memory_format (#37963)

suo committed 4 years ago

4e93844a

Fix splitWithTail to insert the tail immediately after the outer loop. (#37941)

resistor committed 4 years ago

65260d48

[Fakelowp] Open source fake fp16 FC ops (#37923)

Yinghai Lu committed 4 years ago

9143d7fb

Reland [quant][tests] Enable tests to run on all qengine backends (#37943)

supriyar committed 4 years ago

76c964df

[ONNX] Improve error checking for large model export (#37798)

BowenBao committed 4 years ago

122587dc

Report test stats (#37803)

malfet committed 4 years ago

385f7e59

Skip c2_ref_tests on network failures (#37972)

malfet committed 4 years ago

952e0f00

Add option to run python unittests in parallel (#37180)

malfet committed 4 years ago

72e5b7ae

Move complex utilities out of Half.h (#37676)

zasdfgbnm committed 4 years ago

681c6fb6

updated create input and add test methods and added a whitelist for complex (#37835)

anjali411 committed 4 years ago

63428211

Add missing c10::complex::value_type (#37677)

zasdfgbnm committed 4 years ago

14fc83eb

move quantization normalization layers to aten/src/ATen/native/quantized/cpu/ (#37352)

vkuzo committed 4 years ago

09bedec2

add quantized instancenorm operator (#36847)

vkuzo committed 4 years ago

4fa049c5

add quantized groupnorm operator (#36835)

vkuzo committed 4 years ago

b837d5d4

quant: remove hypothesis and int32 from layernorm test (#37947)

vkuzo committed 4 years ago

288dd337

add docker image build ubuntu16.04-cuda9.2-cudnn7-gcc5.4-py3.6 (#37610)

lixinyu committed 4 years ago

675e77e8

Give at::cuda::blas::gemv<at::Half> parity with <float> and <double>. Nature is healing. (#37569)

definitelynotmcarilli committed 4 years ago

35693e9b

[JIT] remove list_with_default op (#37886)

Elias Ellison committed 4 years ago

28ed04c6

Install HugePagesArena to optimize pytorch prediction performance (#37640)

Allan Di Wu committed 4 years ago

f538cd62

Update bazel to 3.1.0 (#37951)

malfet committed 4 years ago

3cc50625

[quant][fix] A typo in quantized::conv2d_relu (#37964)

jerryzh168 committed 4 years ago

56fc347e

Port existing zero_dim_dispatch optimizations from codegen and remove codegen capability. (#37615)

gchanan committed 4 years ago

f29f96d4

Newer Older