pytorch/xla

Pull Requests Commits

alanwaketan committed 1 year ago

c8ad3203

alanwaketan committed 1 year ago

bee14ede

[SPMD] Introduce high level manual sharding APIs (#6931)

alanwaketan committed 1 year ago

Verified 9b2ac4b1

Add examples for training (#6929)

qihqi committed 1 year ago

Verified 781ee938

Add CONTRIBUTING.md (#6924)

qihqi committed 1 year ago

Verified aee1d37f

[SPMD] Support SPMDShardToFullShape (#6925)

alanwaketan committed 1 year ago

Verified 6e34bbed

Update CUDA 12.1 dev image to python 3.10 (#6928)

will-cromar committed 1 year ago

Verified 42b028e2

[SPMD] Fix test_spmd_full_to_shard_shape for v2 (#6927)

alanwaketan committed 1 year ago

Verified dd726d3f

Add support for few basic mutation ops (#6895)

qihqi committed 1 year ago

Verified dc17f9c0

[SPMD] Support SPMDFullToShardShape (#6922)

alanwaketan committed 1 year ago

Verified 2763248f

Add a devcontainer for gpu dev (#6923)

JackCaoG committed 1 year ago

Verified a7a1357d

Add dynamo expand test. (#6659)

ysiraichi committed 1 year ago

Verified 58a412cb

Build plugin during nightly build (#6917)

will-cromar committed 1 year ago

Verified 5a269d7b

Install libtpu for release dockers (#6919)

alanwaketan committed 1 year ago

Verified ec5615c9

[SPMD] Support manual sharding (#6915)

alanwaketan committed 1 year ago

Verified e5513ff4

fix pixel_shuffle return empty (#6907)

zpcore committed 1 year ago

Verified 43f10715

Build TPU package in CI (#6918)

will-cromar committed 1 year ago

Verified 7375a8ad

Add an option to not generate weights. (#6909)

qihqi committed 1 year ago

Verified fc11bbbf

Add `xla.sync` (#6914)

will-cromar committed 1 year ago

Verified 42452717

[Fori_loop|While_loop] Fori loop wrapped from while_loop (#6850)

ManfeiBai committed 1 year ago

Verified f0354ecc

[torch_xla2] Combined tensor.XLADispatchMode() and functions.XLAFunctionMode() (#6902)

will-cromar committed 1 year ago

Verified cfc70e68

Fix runtime error when run dynamo with a profiler scope (#6913)

JackCaoG committed 1 year ago

Verified 7dfbef02

Add ansible stage for building CUDA plugin (#6910)

will-cromar committed 1 year ago

Verified a170ffe6

[Distributed] Switch all_reduce to use the new functional collective op (#6887)

yifuwang committed 1 year ago

Verified a816c42f

fix wheel link (#6908)

zpcore committed 1 year ago

Verified 756b0eca

Update XLA pin to 04/08/2024 (#6904)

JackCaoG committed 1 year ago

Verified fd04a20c

[torch_xla2] Implement some common tensor constructors (#6874)

will-cromar committed 1 year ago

Verified 786484d2

[Pallas] Remove torch.empty in tracing (#6897)

alanwaketan committed 1 year ago

Verified 66ed39ba

Set all_reduce_token to null when exiting (#6898)

yitongh committed 1 year ago

Verified bd2b8fc8

update rc12 for 2.3 release (#6903)

lsy323 committed 1 year ago

Verified 1fe6be30

Older