pytorch/xla

Pull Requests Commits

Trace model hack

asuhan committed 5 years ago

da1642d6

Hack for eager CUDA

asuhan committed 5 years ago

1b1b30cb

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

Remove unused variable from PadInputFromOutputSize (#2456)

asuhan committed 5 years ago

Verified dc2a2256

.circleci: Make the grabbing of the docker tag dynamic (#2436)

seemethere committed 5 years ago

Verified 1f185dc9

Remove use of legacy type() API. (#2454)

ailzhang committed 5 years ago

Verified 47856c8c

Clamp the intermediate value in bce lowering to avoid nan (#2448)

JackCaoG committed 5 years ago

Verified 52eff631

Handle the tuple dtype in pytorch_test_base.py (#2450)

JackCaoG committed 5 years ago

Verified e6d7e57d

Remove old skipped test which was a typo. (#2449)

ailzhang committed 5 years ago

Verified 131bba8b

Add torch.complex32 to list of unsupported types (#2447)

malfet committed 5 years ago

Verified b1e4f936

Skip tests against strides. (#2444)

ailzhang committed 5 years ago

Verified 4c86d878

Fix debug_cloudbuild (#2443)

zcain117 committed 5 years ago

Verified 4df70b27

Update README for 1.6 version. (#2440)

zcain117 committed 5 years ago

Verified bfcf582e

Older