strisunshinewentingwang/pytorch

Pull Requests Commits

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

Exclude more files in torch/csrc/distributed when USE_DISTRIBUTED=0 (#28621)

vishwakftw committed 4 years ago

aea94de0

Explain how to specify library location for MKL (#28779)

adamjstewart committed 4 years ago

4cf7277d

Return None correctly from `Tensor.names` (#28659)

zou3519 committed 4 years ago

5da932ad

addmm: Fix handling of case with empty tensor (#28613)

ggoossen committed 4 years ago

c60dee27

Extend HasElements to support multiple inputs (#28717)

houseroad committed 4 years ago

c89340f0

Eliminate some unnecessary tensor ref count bumps.

Owen Anderson committed 4 years ago

7df3366f

Move mode_t declaration in PadOptions (#28760)

Will Feng committed 4 years ago

f43194ed

Refactor qconv_prepack and qconv_unpack to support conv3d (#28481)

xiaomengy committed 4 years ago

d5afd975

Improve `Tensor` type hints (#28578)

henribru committed 4 years ago

764e0ee8

Type hints: Return `Iterator` instead of `Iterable` from `__iter__` (#27445)

henribru committed 4 years ago

440b1920

Abstract tracer::enter and tracer::exit into a function

James Reed committed 4 years ago

f782500e

Back out D17980308-D17980313 (#28748)

VitalyFedyunin committed 4 years ago

7ff272c6

Automation scripts for perf testing (#28622)

xta0 committed 4 years ago

e96ea288

Support MultiheadedAttention module (#28555)

davidriazati committed 4 years ago

dbf1996f

Newer Older