pytorch/xla

Pull Requests Commits

Mention remote build in CONTRIBUTING.md (#8699)

tengyifei committed 223 days ago

Verified 45af5193

Add missing NeuronPlugin configure_single_process function (#8694)

jeffhataws committed 224 days ago

Verified 065cb5b1

Fix a bug in flash attention where kv_seq_len should divide block_k_major. (#8671)

zhangp365 committed 225 days ago

Verified cff9f4e0

Add another C++11 ABI docker reference (#8649)

tengyifei committed 228 days ago

Verified c0afda3a

[torchax] Update installation guide in readme (#8664)

lsy323 committed 228 days ago

Verified ace5e24b

Fix dk/dv autograd error on TPU flash attention (#8685)

zmelumian972 committed 228 days ago

Verified 0cd1fc2a

Added pow to autocast policy and unit test (#8684)

avizon-aws committed 228 days ago

Verified afaf0d05

Misc fixes (#8679)

qihqi committed 229 days ago

Verified 8b45e599

Fix aten.linear and add aten.einsum (#8666)

yaochengji committed 229 days ago

Verified 45109d1e

fixed the type miss match issue with symint in SPMD regime and dynamo (#8681)

hosseinsarshar committed 229 days ago

Verified b9bef193

fix a bug in tensor.flatten (#8680)

yaochengji committed 230 days ago

Verified 49cd3105

Install `clang-17` on development Dockerfile. (#8673)

ysiraichi committed 231 days ago

Verified ec10537c

[CI] Fix torchbench setup failure (#8670)

lsy323 committed 231 days ago

Verified a4356250

Add ragged paged attention (#8659)

vanbasten23 committed 231 days ago

Verified 84800944

`cummax`: fix 0-sized dimension reduction. (#8653)

ysiraichi committed 231 days ago

Verified 1c537904

[torchax][distributed] Use local device for default jax device (#8667)

lsy323 committed 232 days ago

Verified 35789406

Dynamo/AOTAutograd traceable flash attention (#8654)

zpcore committed 234 days ago

Verified 9ae017e2

openxla pin update to 20250131 (#8621)

zpcore committed 235 days ago

Verified 82d35047

[torchax] fix einsum op conversion (#8657)

yaochengji committed 235 days ago

Verified 8572e750

Fix the retention for existing gradients in the Grad Acc API (#8658)

rpsilva-aws committed 235 days ago

Verified 39dd7959

Add a stub for host offloading docs (#8656)

tengyifei committed 236 days ago

Verified e583c2cc

Fix C++11 ABI wheel instructions (#8650)

tengyifei committed 236 days ago

Verified 93a2ba6b

Update README.md for PyTorch/XLA 2.6 release (#8647)

tengyifei committed 237 days ago

Verified d9593f03

Fix metrics saver when a file location is specified (#8644)

yaochengji committed 237 days ago

Verified 7243b808

Build v2.6 release wheel (#8645)

tengyifei committed 237 days ago

Verified 01839974

Add a lowering to linear instead of using torch decomp (#8643)

qihqi committed 237 days ago

Verified 25d09fdd

Prepare v2.6.0-rc10 (#8639)

tengyifei committed 238 days ago

Verified 310d2d81

scan documentation (#8631)

tengyifei committed 238 days ago

Verified 225c65bd

Fix grad accumulation tests on TPUv5/v6 (#8628)

tengyifei committed 239 days ago

Verified 8e6ca600

Rename torch_xla2 to torchax (#8599)

qihqi committed 241 days ago

Verified 8b241409

Newer Older