pytorch/xla

Pull Requests Commits

Move tensors to XLA

changm committed 1 year ago

Verified a014fe08

Replicate test dynamic shapes with torch.compile interface

changm committed 1 year ago

Verified bb740d42

Update OpenXLA pin to fix GPU build (#6569)

cota committed 1 year ago

Verified 2bbc9a49

[Fori_loop|While_loop] Placeholder lower torch.while_loop with python dispatch for simple addition test case (#6532)

ManfeiBai committed 1 year ago

Verified 5405c4d4

Enable graph input matcher test (#6591)

JackCaoG committed 1 year ago

Verified 69d453ea

Add extra APIs. (#6586)

qihqi committed 1 year ago

Verified 5145ea1c

Add XLA specific XPlane profiling. (#6552)

golechwierowicz committed 1 year ago

Verified adabb59a

Fix some more core aten ops (#6580)

wonjoo-wj committed 1 year ago

Verified 2cb31440

fix linter (#6585)

JackCaoG committed 1 year ago

Verified c524b4dc

Upgrade C++ formater to clang-format-11 (#6583)

JackCaoG committed 1 year ago

Verified 0c3fbc66

Register OpOverload when OpOverloadPacket is registered (#6579)

lsy323 committed 1 year ago

Verified d8ed46d5

Add basic device APIs to the top-level `torch_xla` module. (#6571)

will-cromar committed 1 year ago

Verified 0ec5b917

Fix typo in TROUBLESHOTTING (#6468)

JackCaoG committed 1 year ago

Verified e77a629c

Fix a incorrect assert on frame count for PT_XLA_DEBUG=1 (#6466)

JackCaoG committed 1 year ago

Verified c08ae21a

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

Older