pytorch/xla

Pull Requests Commits

alanwaketan committed 2 years ago

743978fe

Make custom_sharding_

alanwaketan committed 2 years ago

10b07b70

alanwaketan committed 2 years ago

2cb25577

Revert vscode init

alanwaketan committed 2 years ago

97201827

alanwaketan committed 2 years ago

7c57ee1c

alanwaketan committed 2 years ago

6cbf580f

Do custom call in _propagate_xla_data

alanwaketan committed 2 years ago

605b46e8

alanwaketan committed 2 years ago

3c775212

alanwaketan committed 2 years ago

63a5734d

Make it in-place op

alanwaketan committed 2 years ago

da528248

Don't use XLAShardedTensor

alanwaketan committed 2 years ago

e45ec012

alanwaketan committed 2 years ago

2d468d52

Fix the custom call hlo

alanwaketan committed 2 years ago

c4ff7d92

Modify test and small fix

alanwaketan committed 2 years ago

1415d565

connect the pybind

alanwaketan committed 2 years ago

9622c475

add tensor methods

alanwaketan committed 2 years ago

ac26dafa

alanwaketan committed 2 years ago

1ba36186

alanwaketan committed 2 years ago

43047804

alanwaketan committed 2 years ago

5459ca17

alanwaketan committed 2 years ago

1036b93f

[SPMD] Fix HybridMesh ordering (#5478)

alanwaketan committed 2 years ago

Verified b8b48451

Don't move full tensor to device in deferred_init (#4819)

Zach Zheng committed 2 years ago

Verified c57370dc

Enable PJRT C API Client and other changes for Neuron (#5428)

aws-kingrj committed 2 years ago

Verified daed1d0a

Introduce `torch_xla.runtime.use_spmd()` (#5474)

yeounoh committed 2 years ago

Verified 462cb0f4

Enabling sharding device data IR (#5475)

JackCaoG committed 2 years ago

Verified 9a5fdf39

Fix the input sharding for dynamo (#5469)

JackCaoG committed 2 years ago

Verified bf997819

bump the timeout for CI (#5470)

JackCaoG committed 2 years ago

Verified 76495f0b

Fix data type in Pow with Scalar base and Tensor exponent (#5467)

lsy323 committed 2 years ago

Verified 909f28fa

update remote cache key in ansible (#5463)

lsy323 committed 2 years ago

Verified 028df4da

[GPU]Update README.md with wheel/docker for CUDA12.0 and deprecate CUDA11.7 (#5443)

ManfeiBai committed 2 years ago

Verified 95708f2b

Older