pytorch/xla

Pull Requests Commits

lower replication_pad3d and replication_pad3d_backward (#6566)

ManfeiBai committed 1 year ago

Verified 0192ff75

Update export.py (#6576)

wang2yn84 committed 1 year ago

Verified 8c65f099

Adds complex64 conversion between Jax and Pytorch to tensor.py (#6575)

wang2yn84 committed 1 year ago

Verified ee341ee7

Do not cache input args in dynamo bridge (#6553)

lsy323 committed 1 year ago

Verified 6aeab300

Add more pjrt runtime tests for GPU. (#6514)

vanbasten23 committed 1 year ago

Verified 19362a98

Correct a typo in TROUBLESHOOTING.md (#6534)

vanbasten23 committed 1 year ago

Verified 17541938

[benchmarks] Update `DENY_LIST`. (#6558)

ysiraichi committed 1 year ago

Verified f4971a7d

[benchmarks] Fix AMP data-type. (#6550)

ysiraichi committed 1 year ago

Verified 3d21839d

Add fallback check to test_core_aten_ops.py (#6559)

wonjoo-wj committed 1 year ago

Verified 423b9a30

Add scalar_tensor op; add more tests. (#6549)

qihqi committed 1 year ago

Verified ed71f922

Add default flag `xla_tpu_prefer_async_allgather_to_allreduce` (#6528)

jonb377 committed 1 year ago

Verified 07f68a0f

Revert "[Core Aten Op] bring aten_replication_pad3d test back" (#6554)

wonjoo-wj committed 1 year ago

Verified c20ebd79

Update contribution docs to explain how to update libtpu (#6551)

changm committed 1 year ago

Verified f4e2f776

Enable tests for aten.any* (#6545)

qihqi committed 1 year ago

Verified a7a26917

Disable GPU coverage upload on the master (#6543)

yeounoh committed 1 year ago

Verified e47810f4

[benchmarks] Default to `bfloat16` (inference) and AMP (training) precision. (#6518)

ysiraichi committed 1 year ago

Verified 20692cb0

[Core Aten Op] bring aten_replication_pad3d test back (#6537)

ManfeiBai committed 1 year ago

Verified b7c760da

Dont wrap ints in torch.tensor; (#6539)

qihqi committed 1 year ago

Verified 911485e7

Add limit to max inflight TPU computations (#6533)

will-cromar committed 1 year ago

Verified 29b01c6a

OpenXLA pin update (#6530)

yeounoh committed 1 year ago

Verified 27d1f703

Update TPU CI workflow to trigger on push to master and re-enable spm… (#6531)

mbzomowski committed 1 year ago

Verified 1d4a8da5

[FSDPv2] Move the module to xla device (#6525)

alanwaketan committed 1 year ago

Verified afc1f0e2

Modify export to use refactored _graph_module_flat_inputs (#6526)

qihqi committed 1 year ago

Verified 959e4786

aggregate.py: fix detection of no datapoints (#6523)

cota committed 1 year ago

Verified d6bc4f31

aggregate.py: fix relative error computation (#6522)

cota committed 1 year ago

Verified 7da7ee42

Add CUDA as supported xla backends to github issue template (#6517)

vanbasten23 committed 1 year ago

Verified e736de87

Misc fix for jax export, ran formatter (#6516)

qihqi committed 1 year ago

Verified 1bcc0145

Lower the embedding op (#6495)

bhavya01 committed 1 year ago

Verified 657b6925

Lowering for _pdist_forward (#6507)

zpcore committed 1 year ago

Verified a5856f95

Enable tests for reflection_pad1d, reflection_pad3d, grid_sampler_2d (#6513)

qihqi committed 1 year ago

Verified 2797852a

Newer Older