pytorch/xla

Pull Requests Commits

Restructure setup.py and add the GIT revision at which the library was built.

dlibenzi committed 6 years ago

bec9d364

Change to per-file import

Ailing Zhang committed 6 years ago

da9ae4c5

Move torch_xla_py to torch_xla. Fix all uses

Ailing Zhang committed 6 years ago

3f3bfa04

Use remote handle value to make parameters unique and added check to verify that supplied parameters are matching the XLA program shape.

dlibenzi committed 6 years ago

ed502f69

Better stack trace argument dedup. (#1082)

dlibenzi committed 6 years ago

1ecceec0

Add slow XLA compile tracking and more graph hash logging. (#1078)

dlibenzi committed 6 years ago

8523c0df

Made XLA to() API support self being a CPU tensor. (#1079)

dlibenzi committed 6 years ago

71fb134a

Disable two failed tests from multiple dispatch.

Ailing Zhang committed 6 years ago

c58e186c

Fix up comment to clarify only TPU Pods be used (#1076)

jysohn23 committed 6 years ago

Verified 3e12428e

Use docker version 347. (#1071)

ailzhang committed 6 years ago

Verified 4e5dc58b

Pods script placement (#1074)

jysohn23 committed 6 years ago

Verified a408e77d

Unbreak tests - importing common_utils fails (#1073)

jysohn23 committed 6 years ago

Verified 0adb77ef

Switch enable_whole_mesh_comp on clusters jobs too

jysohn23 committed 6 years ago

8744b416

Do not fiddle with syspath.

dlibenzi committed 6 years ago

0534e0d2

Makes allowed tests a non-diff list, updates diff to handle dtypes (#1069)

mruberry committed 6 years ago

Verified 4f4ddcb5

Diagonal as view.

dlibenzi committed 6 years ago

040ad82e

Only override pytorch/xla sys.path when required (#1066)

jysohn23 committed 6 years ago

Verified dd6493b1

Updating the API guide with a few more pitfalls. (#1059)

taylanbil committed 6 years ago

Verified 9be6d100

Give backward() a value to start computation. (#1065)

ailzhang committed 6 years ago

Verified 2d20994f

Set the default device for C++ tests.

dlibenzi committed 6 years ago

6fd97774

Fix potential missed-free race in XRT handle.

dlibenzi committed 6 years ago

458c76d2

Enable full mesh compile setup.

dlibenzi committed 6 years ago

63f5f0f6

Add pytorch-tpu/examples to the docker image. (#1058)

taylanbil committed 6 years ago

Verified 3ee85d75

Add test_clamp to allowed list. (#1062)

ailzhang committed 6 years ago

Verified c117cdff

Add torch_xla/version.py to .gitignore

jysohn23 committed 6 years ago

37179291

Add torch_xla.__version__

jysohn23 committed 6 years ago

afddaf0b

added master_print function (#1060)

taylanbil committed 6 years ago

Verified 050ed2b1

Temporarily disable pinverse test (#1057)

ailzhang committed 6 years ago

Verified 1a1fa2b4

ezyang committed 6 years ago

2b90f961

ezyang committed 6 years ago

6ae60def

Older