pytorch/xla

Pull Requests Commits

Fix failing tests

JackCaoG committed 2 years ago

6632e480

Check before doing mark_step for dynamo

JackCaoG committed 2 years ago

82164f05

Optimize Dynamo Training Execute

JackCaoG committed 2 years ago

30e78848

Enable python 3.8 and resolve dependency issues (#4529)

yeounoh committed 2 years ago

Verified bfd20f37

Try bootstrapping tpu env from env vars (#4499)

steventk-g committed 2 years ago

Verified f2cb76f9

Move XLA testing to Python-3.8 (#4527)

malfet committed 2 years ago

Verified 4f4ab8a3

Increase the timeout to make cuda docker build more stable (#4522)

JackCaoG committed 2 years ago

Verified b6b4da0d

[PJRT] Release the GIL during `TransferFromServer` (#4504)

will-cromar committed 2 years ago

Verified d41c5ba2

Always use `ExecuteSharded` future with PJRT C API. (#4521)

will-cromar committed 2 years ago

Verified efb53533

Disable the Sandbox build in Docker (#4515)

JackCaoG committed 2 years ago

Verified 661ab6d2

Bump up the xla_base image ver to v0.8 in upstream repo (#4516)

yeounoh committed 2 years ago

Verified 6acc0c9c

Switch to pytorch sccache (#4489)

huydhn committed 2 years ago

Verified 6bf36302

Remove deprecated torch.symeig (#4498)

IvanYashchuk committed 2 years ago

Verified 903f2ac5

Update the tf pin to 2023 01/23 (#4474)

JackCaoG committed 2 years ago

Verified 2baef35b

Don't use futures on output buffers (#4503)

will-cromar committed 2 years ago

Verified b5d6be81

Fix HLO dump file names for PJRT (#4487)

will-cromar committed 2 years ago

Verified b9efeeb4

Disable nonzero test to unblock the head (#4512)

JackCaoG committed 2 years ago

Verified 5714e03f

Disable more dynamic shape tests (#4507)

JackCaoG committed 2 years ago

Verified 05195c7e

Temporarily disable test TestDynamicShape.test_nonzero_cast (#4502)

vanbasten23 committed 2 years ago

Verified bc14fe02

refactor: move dynamo/TorchXLA bridge to pytorch/xla repo (#4476)

shunting314 committed 2 years ago

Verified 0c9801bc

Revert "Squashed commit of the following: (#3272)" (#4497)

JackCaoG committed 2 years ago

Verified 844f137c

Squashed commit of the following: (#3272)

IvanYashchuk committed 2 years ago

Verified 282567af

Forward fix for 81763 (#4490)

lezcano committed 2 years ago

Verified 70cc6295

Revert "Revert "Adapt to new pytorch grad clipping API (#4480)" (#4484)" (#4486)

wonjoo-wj committed 2 years ago

Verified aff35132

Revert "Adapt to new pytorch grad clipping API (#4480)" (#4484)

wonjoo-wj committed 2 years ago

Verified 3e6bbbbd

Adapt to new pytorch grad clipping API (#4480)

wonjoo-wj committed 2 years ago

Verified eac4e547

Companion PR for https://github.com/pytorch/pytorch/pull/92107 (#4463)

ezyang committed 2 years ago

Verified 5c54655a

Fix shm size (#4470)

steventk-g committed 2 years ago

Verified d636e777

Code-gen LeakyRelu, LeakyReluBackward again (#4468)

lsy323 committed 2 years ago

Verified ba61212b

Enable lowering for upsample_bilinear2d with scale factor (#4464)

lsy323 committed 2 years ago

Verified d74faafc

Older