strisunshinewentingwang/pytorch

Pull Requests Commits

Improve TORCH_LOGS settings error msg (#97264)

mlazos committed 1 year ago

c37ab85d

inductor(cpu): support mkldnn packed linear to improve bfloat16 performance (#96954)

XiaobingSuper committed 1 year ago

aab34a47

Changed logging in aotautograd a little (#97289)

Chillee committed 1 year ago

e49b4d38

Enhance error message for dependency check (#96642)

wangxiyuan committed 1 year ago

4ab1588d

Added current buck target as minifier dep (#97183)

mlazos committed 1 year ago

f6bafcde

Init quantization backend config for inductor (#96476)

leslie-fang-intel committed 1 year ago

a6d8c709

[Inductor] Enable CppWrapper to support BF16 (#97089)

EikanWang committed 1 year ago

517a432d

[Inductor] Fix the issue that cannot pass lint check for debug mode (#97249)

EikanWang committed 1 year ago

573b2deb

[Inductor] Load a BF16 scalar and broadcast it as a float vector (#97070)

EikanWang committed 1 year ago

37e1d858

[Inductor] Fix the issue that cannot pass lint check for debug mode (#97249)

EikanWang committed 1 year ago

c5d7ed94

Move empty check to the start of _pack_padded_sequence (#94885)

CaoE committed 1 year ago

b72bddab

Remove chhillee from autoreview (#97293)

Chillee committed 1 year ago

f9a9a888

Update NestedTensor add to support non identical striding for NT+NT (#97195)

drisspg committed 1 year ago

db15d191

[Inductor] Add debug logging to explain reasons of disabling vectorization (#97108)

jgong5 committed 1 year ago

4733de18

[Dynamo] throw better error message if assert with non-string message (#97297)

yanboliang committed 1 year ago

c1025af0

Test and fix guard fail message in CompileProfiler (#97055)

wconstab committed 1 year ago

57c13fde

Mention pytorchbot command on label error (#97267)

kit1980 committed 1 year ago

1e4e2567

Add sympy to binary linux test - fix conda nightly (#97281)

atalman committed 1 year ago

688427b5

[Dynamo] Support nn.Module.named_children (#97216)

yanboliang committed 1 year ago

c7fad133

DCE inference graphs too (#97275)

ezyang committed 1 year ago

aa3a57b8

[inductor] reusing autotuning sub-processes (#97219)

shunting314 committed 1 year ago

3282030f

Add gradcheck_nondet_tol to a few padding moduleinfos (#97265)

mikaylagawarecki committed 1 year ago

0b094ca3

[draft for discussion] add per-dispatch key modes (#97052)

bdhirsh committed 1 year ago

af440c42

Fix dispatching issue of the new device type. (#97273)

egienvalue committed 1 year ago

793cf0cb

moving nvfuser benchmark to third_party/nvfuser (#96725)

jjsjann123 committed 1 year ago

2b32a74a

[BE] Remove spurious semicolon in XPUHooksInterface.h (#97296)

malfet committed 1 year ago

a1ef0be3

Fixes warning to refer to SMs instead of Cuda Cores (#97224)

Chillee committed 1 year ago

6dded5d6

leave libdevice name for fbcode (#97257)

ngimel committed 1 year ago

47f18b78

Fix kDefaultTimeout multiple definition build failure (#97270)

sclarkson committed 1 year ago

9a189682

[inductor] hoist symbolic padding expressions (#97099)

ngimel committed 1 year ago

e7d93316

Newer Older