pytorch/xla

Pull Requests Commits

JackCaoG committed 1 year ago

eecda39e

Optimize execution for ops that have multiple output in eager mode

JackCaoG committed 1 year ago

e36f0da6

support gmm as a custom op for dynamo (#7672)

JackCaoG committed 1 year ago

Verified 1651e764

Update openxla pin to 11 July 2024 (#7639)

bhavya01 committed 1 year ago

Verified 8043050f

Replace remaining usages of `jax.lax.Precision` (#7671)

will-cromar committed 1 year ago

Verified ac26a977

Add shard_map wrapper (#7670)

lsy323 committed 1 year ago

Verified 18ed9dd8

[Fori_loop] Update `randint` max range to Support `bool` dtype (#7632)

huzama committed 1 year ago

Verified 41d998de

Use jax.export insteadl of jax.experimental.export (#7300)

qihqi committed 1 year ago

Verified 34736f0f

lower _aten_constant_pad_nd (#7613)

wenxindongwork committed 1 year ago

Verified b32a91f2

Add `vsplit`, `vstack`, `vdot` ops (#7664)

miladm committed 1 year ago

Verified 66ff604a

[torch_xla2] Functional collectives (#7580)

will-cromar committed 1 year ago

Verified 3078d163

Asymmetric quantized matmul support (#7626)

lsy323 committed 1 year ago

Verified 289471cb

Easier way to make img show up in our doc page (#7644)

JackCaoG committed 1 year ago

Verified 72069750

Move tensor constructors to `jaten` (#7615)

will-cromar committed 1 year ago

Verified 5e9da938

[torch_xla2] Add broadcast_to, isin, item, lcm, lerp (#7638)

ManfeiBai committed 1 year ago

Verified ccd3da2c

[torch_xla2] Add fmax, fmin, frexp (#7603)

ManfeiBai committed 1 year ago

Verified c62b1d46

Add blockwise quantized dot support (#7605)

lsy323 committed 1 year ago

Verified 88bcb45f

Update `apt-get update` in upstream image build (#7640)

will-cromar committed 1 year ago

Verified edd24532

separate PyTorch/XLA docs (#7593)

JackCaoG committed 1 year ago

Verified 275bf919

Update test_pallas_spmd.py for CI (#7637)

bhavya01 committed 1 year ago

Verified c84c8932

Fallback `_embedding_bag_backward` and force `sparse=false`. (#7584)

ysiraichi committed 1 year ago

Verified 3bcb1fb1

fix eager mode spmd module loading with fsdpv2 (#7631)

JackCaoG committed 1 year ago

Verified 10a51300

Update test_pallas.py for CI (#7629)

ManfeiBai committed 1 year ago

Verified 05a6487f

Introduce CUDA OpenXLA fallback. (#7318)

ysiraichi committed 1 year ago

Verified c782e0de

Early exit for clip_grad_norm_ (#7621)

alanwaketan committed 1 year ago

Verified 172097b6

[Pallas] Improve the user guide (#7625)

alanwaketan committed 1 year ago

Verified 661f1914

SPMD parameter wrapping fix review comment (#7619)

JackCaoG committed 1 year ago

Verified 3ce6ffb2

add tpu example for torchrun on GKE (#7620)

chengcongdu committed 1 year ago

Verified 51f89c85

Update nightly package version to 2.5 (#7623)

bhavya01 committed 1 year ago

Verified bbf0db79

added how to tell when ops need to be lowered (#7624)

nhira committed 1 year ago

Verified 2c011379

Older