strisunshinewentingwang/pytorch

Pull Requests Commits

Fix build error in VariableTypeManual

linbinyu committed 4 years ago

1e3e1f5b

Updating submodules

svcscm committed 4 years ago

c6ad68cf

Add the unittest import for test_fake_quant.py (#28815)

jianyuh committed 4 years ago

6f90567e

Small fixes for torchbind (#28800)

James Reed committed 4 years ago

949678bd

Update C++ parity table for torch::nn::Linear (#28804)

Will Feng committed 4 years ago

f63cf96c

Simplify copy kernel (#28428)

zasdfgbnm committed 4 years ago

5c5b2c68

Make TensorIterator stop promoting types by copying (#28427)

zasdfgbnm committed 4 years ago

b9f099ed

Move type casting to c10/util/TypeCast.h (#28426)

zasdfgbnm committed 4 years ago

688a9dbe

Return NotImplemented from all binary math ops (#27423)

peterbell10 committed 4 years ago

f33813d5

Add the warning message for API with linear modules (#28766)

jianyuh committed 4 years ago

9e64c54c

Temporarily disable test_numerical_consistency_per_channel due to failure (#28807)

jianyuh committed 4 years ago

02d31846

make sure clang-tidy is diffing against the right thing (#28788)

suo committed 4 years ago

9f890a92

Deprecate torch::nn::modules_ordered_dict API (#28774)

Will Feng committed 4 years ago

5804e54c

Back out "Add memory format support to `full_like` operator" (#28803)

VitalyFedyunin committed 4 years ago

87c98acf

Back out "Add memory format support to `ones_like` operator" (#28802)

VitalyFedyunin committed 4 years ago

d828fef8

Back out "Add memory format support to `rand_like` operator" (#28801)

VitalyFedyunin committed 4 years ago

266c1652

C++ API: torch::nn::LPPool2d (#28492)

nuka137 committed 4 years ago

648749b2

Enabling intra-op parallelism for dynamic quantized Linear operator (#28477)

jianyuh committed 4 years ago

052046b1

separate PT and C2 to reduce build time (#28731)

mingzhe09088 committed 4 years ago

9f44a046

Fix obviously-broken .clang-tidy files (#28547)

wfarner committed 4 years ago

0c7537c4

Reduce logging frequency for empty range tolarence

Xinyi Zhang committed 4 years ago

f5ea2ca3

Change reorder_dimensions behavior to favor output writting sequence (#28615)

Xinyu Li committed 4 years ago

7ed9a3ec

Remove the redundant calculation of derivative of power function (#28651)

titaneric committed 4 years ago

82f31e02

Added docs for context method mixins. Fixes issue #27365 (#28643)

prasunanand committed 4 years ago

4230132b

Updating submodules

svcscm committed 4 years ago

0e86c99b

provide memory format as Contiguous explicitly when calling to clone() (#28029)

ifedan committed 4 years ago

5835ad07

Kill _th_zero binding, just use a simple native function instead.

gchanan committed 4 years ago

6eaea398

Kill THTensor_(match), which isn't used.

gchanan committed 4 years ago

8e67b78d

Write out some set_ overloads instead of relying on code binding generation.

gchanan committed 4 years ago

1e5b2559

adding python all_gather coalesced functionality and testing. (#28634)

agolynski committed 4 years ago

45dab561

Newer Older