pytorch/xla

Pull Requests Commits

Try adding git config before running git revert.

zcain117 committed 5 years ago

2c088655

Add in the revert.

zcain117 committed 5 years ago

14c8c487

Rebased. Not using git revert.

zcain117 committed 5 years ago

e7891156

Remove the git revert. Still have conda install dataclasses.

zcain117 committed 5 years ago

9f560eee

Use aug 7 but use git revert 18fa011c

zcain117 committed 5 years ago

a697cc0f

Use Aug7 commit for both xla and pytorch.

zcain117 committed 5 years ago

774d2218

Use Aug 4 pytorch along with same partial-day Aug 3 pytorch/xla

zcain117 committed 5 years ago

3bcaff33

Exclude 'Disable CXX 11 ABI when building wheels (#2390)' but keep the other commit from Aug 3 on the pytorch/xla side

zcain117 committed 5 years ago

2b598b7f

Use Aug 4 xla and Aug 3 pytorch

zcain117 committed 5 years ago

a5a7ebea

Use aug 3 version for both pytorch/xla and pytorch

zcain117 committed 5 years ago

8ab46cff

Add ability to check out xla and pytorch at different commits

zcain117 committed 5 years ago

2c868806

Simple setup docs and package name override to avoid conflicts when necessary (#2481)

cjolivier01 committed 5 years ago

Verified 8af57fb5

Only overwrite test precision if the pt/xla tolerance is larger (#2478)

JackCaoG committed 5 years ago

Verified 8b9f5c29

Update doc to build docker inside xla directory (#2474)

JackCaoG committed 5 years ago

Verified a2a53222

Remove float16 test from skip test since they should get auto skip (#2475)

JackCaoG committed 5 years ago

Verified 006cf17f

Remove registering fallthrough (moving to pytorch core). (#2463)

ailzhang committed 5 years ago

Verified c6082574

Cache debug docker and new conda install (#2476)

zcain117 committed 5 years ago

Verified d214e464

Add all_to_all back to our doc (#2472)

JackCaoG committed 5 years ago

Verified df0a2776

Implement the lowering of baddbmm op (#2471)

JackCaoG committed 5 years ago

Verified d2330676

Allow forwarding args to data loader and train/test function + allow "-" and "." in node name (#2470)

Chris Olivier committed 5 years ago

Verified ec79c5f4

Update tensorflow to include the all_to_all fix (#2469)

JackCaoG committed 5 years ago

Verified c712c93e

Make ArgMax and ArgMin return the first index when there’s same values (#2419)

JackCaoG committed 5 years ago

Verified a93682f6

Disable test_bfloat16_xla due to half support (#2466)

JackCaoG committed 5 years ago

Verified 19fc8124

revert empty() signature to TensorOptions version (#2465)

Basil Hosmer committed 5 years ago

Verified 92454efc

Dispatch aten::maximum and aten::minimum to XLATensor::max and XLATensor::min (#2438)

JackCaoG committed 5 years ago

Verified bde3f5c8

Skip memory overlap checking tests (#2464)

peterbell10 committed 5 years ago

Verified c1a6a38a

Add op lowering guide (#2458)

JackCaoG committed 5 years ago

Verified 2e629343

Add a timestamp tag to the debug docker image. (#2462)

zcain117 committed 5 years ago

Verified 29d025fb

Disable test_matrix_exp_analytic_xla on TPU due to server side crash (#2459)

JackCaoG committed 5 years ago

Verified afbc41b1

Lower `aten::round` operator (#2457)

jysohn23 committed 5 years ago

Verified 48911054

Older