pytorch/xla

Pull Requests Commits

Add set_buffer_donor as a pytorch custom op

JackCaoG committed 1 year ago

9f7d2463

Use pybind to set should_alias_with_buffer_donor_config

JackCaoG committed 1 year ago

055b55c8

use contextmanager

JackCaoG committed 1 year ago

ecc426c7

AliasWithBufferDonorContext wtih default value

JackCaoG committed 1 year ago

d9fdbcbe

Add API to donate input buffer for dynamo execution (#6587)

JackCaoG committed 1 year ago

Verified 3e2a23ca

Shorten ResNet tests and add more test coverage. (#6606)

vanbasten23 committed 1 year ago

Verified 65f54651

Move TPU CI into shared script (#6602)

will-cromar committed 1 year ago

Verified bc3615c8

Update nightly to 2.3 (#6617)

JackCaoG committed 1 year ago

Verified 9b81acc0

nightly.sh: build with local compute capability (#6608)

cota committed 1 year ago

Verified 4327d24a

nightly.sh: fix assignment of timestamp default (#6601)

cota committed 1 year ago

Verified a9a788d5

Add ResNet18 per channel quant test, fix assertion in quantize func and name mangling for tf saved model (#6592)

lsy323 committed 1 year ago

Verified 1654ae7f

Fix instruction for getting HLO/IR dump (#6599)

vanbasten23 committed 1 year ago

Verified b219da81

Add topological sort after composite grouping pass (#6607)

lsy323 committed 1 year ago

Verified dc08cb68

[benchmarks] Set autocast kwargs only if AMP. (#6612)

ysiraichi committed 1 year ago

Verified f65bae4d

Conditional run tests that have TF dependency (#6573)

lsy323 committed 1 year ago

Verified 8217c6b5

Make sure torch_xla2 works with torch stable 2.2 (#6593)

qihqi committed 1 year ago

Verified a659d970

[benchmarks] Fix YAML file name. (#6597)

ysiraichi committed 1 year ago

Verified 1a2ea9de

[Core Aten Ops] Lower reflection_pad1d, reflection_pad1d_backward, reflection_pad3d and reflection_pad3d_backward (#6588)

ManfeiBai committed 1 year ago

Verified 13e8647c

Allow Inductor benchmarks for CPU backends. (#6555)

changm committed 1 year ago

Verified 263da4ee

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

Older