strisunshinewentingwang/pytorch

Pull Requests Commits

Avoid mat1 references in mm_mat1_backward (#45777)

kurtamohler committed 3 years ago

28f8372b

caffe2/plan_executor: propagate exceptions from reporter substeps (#46424)

d4l3k committed 3 years ago

dd169ca1

[fx] allow custom behavior for args, kwargs, and bool (#45193)

zdevito committed 3 years ago

24ca2763

[NCCL] Modularize ncclCommWatchdog (#46051)

osalpekar committed 3 years ago

2e2fe8cf

Fix implicit cast in custom_function (#46445)

r-barnes committed 3 years ago

be0c4318

Make `torch.lu` support complex input on CUDA. (#45898)

nikitaved committed 3 years ago

9300a277

[Flaky tests] Fix test_all_gather_timeout test (#45989)

rohan-varma committed 3 years ago

5c5484c8

[caffe2] add concat benchmark (#46457)

jspark1105 committed 3 years ago

c37baa91

Fix out-of-bounds access for caching allocator calls (#46439)

malfet committed 3 years ago

b5702e23

install lcov in Docker image if coverage is specified (#46404)

janeyx99 committed 3 years ago

d6e60730

Fix deprecated warnings for nan_to_num (#46309)

RockingJavaBean committed 3 years ago

7b788d11

[mlf][efficiency] modify equalization scale operator to return single output (#46449)

hwangjeff committed 3 years ago

ecf63351

[TensorExpr] Correctly handle negative dimensions in aten::cat when lowering to tensor expressions. (#46446)

Mikhail Zolotukhin committed 3 years ago

4359c5e0

Remove variable_excluded_from_dispatch() check for factory functions. (#46371)

Ailing Zhang committed 3 years ago

ec5f81f9

fix type check for torch.quantization._numeric_suite (#46330)

Rong Rong committed 3 years ago

d1745c36

Add named tuple's error message and workaround for RET failure (#46347)

jwpark1985 committed 3 years ago

92921c82

Update VariableTypeManual.cpp to not use catchAllKernel. (#46353)

Ailing Zhang committed 3 years ago

d278e83e

Fixes bug in sspaddmm (#45113) (#45963)

vfdev-5 committed 3 years ago

dc7cd974

More Timer refinement (#46023)

Taylor Robie committed 3 years ago

dda95e69

Add Sigmoid operator from Caffe2 (#46286)

b-koopman committed 3 years ago

757173a4

[caffe2] Bypass memonger for in-place ops (#46378)

Hao Lu committed 3 years ago

16c52d91

Update default ouput extension in optimize_for_mobile.cc (#45598)

dreiss committed 3 years ago

faf03bd2

vmap: added fallback for in-place operators (#46191)

zou3519 committed 3 years ago

f96cb9de

Rename createLevelsBitset -> createVmapLevelsBitset; move it (#46190)

zou3519 committed 3 years ago

401c85b4

Add quick fix for view/inplace issue with DDP (#46406)

albanD committed 3 years ago

849bc77e

Remove codegen for old RegistrationDeclarations.h (#46370)

Ailing Zhang committed 3 years ago

ba92920a

Add operator benchmark for 4bit/8bit embedding lookups

bugra committed 3 years ago

03c7d5be

Fixing pow for special case between cuda tensors and cpu tensors and reframed test cases a tiny bit (#46320)

janeyx99 committed 3 years ago

c99378af

Migrate CPU_tensor_apply to TensorIterator (#44242)

muthuArivoli committed 3 years ago

7d6d5f4b

use @mode/ndk_libcxx instead of mode/gnustl

linbinyu committed 3 years ago

8f61fa65

Newer Older