pytorch/xla

Pull Requests Commits

re-annotate _scatter test as static

bfolie committed 274 days ago

Verified 2fdf36ce

Merge branch 'bfolie/all-to-all' of https://github.com/pytorch/xla into bfolie/all-to-all

bfolie committed 274 days ago

eeb45354

Merge branch 'master' of https://github.com/pytorch/xla into bfolie/all-to-all

bfolie committed 274 days ago

09ec6e11

Add r2.9 release nightly tests to README (#9478)

pgmoka committed 274 days ago

Verified 8b95c5d6

Fix spmd sharding visualization when device index is >= 10 (#9475)

jeffhataws committed 275 days ago

Verified c6aa90f1

Update Setuptools version in build dependencies. (#9487)

bhavya01 committed 275 days ago

Verified b612613d

Fix duplicate labels and other docs build warnings (#9446)

melissawm committed 275 days ago

Verified d7c9e506

Bump version (#9480)

qihqi committed 275 days ago

Verified 4b3d34da

Error Handling: refactor the existing `ComputationClient` implementations to use status QOL functions. (#9420)

ysiraichi committed 275 days ago

Verified 866fd22d

Optimize w8a8 pallas kernel (#9473)

kyuyeunk committed 276 days ago

Verified 36ff641d

Torchax: Allow reuse of the jittable procedure with different functio… (#9374)

zmelumian972 committed 276 days ago

Verified ebf9a8be

Update TPU CI with latest docker container (#9483)

bhavya01 committed 276 days ago

Verified ca04583b

Add dtensor mesh conversion test (#9474)

aws-cph committed 276 days ago

Verified 93a5e583

Merge branch 'master' into bfolie/all-to-all

bfolie committed 277 days ago

Verified d143d9ac

implement collective reduce op (#9437)

bfolie committed 277 days ago

Verified 173a7708

Error Handling: replace `ConsumeValue` with `GetValueOrThrow`. (#9464)

ysiraichi committed 279 days ago

Verified 51518f9b

[Kernel] Update ragged attention block table (#9476)

yaochengji committed 280 days ago

Verified 992e87ad

Merge branch 'master' of https://github.com/pytorch/xla into bfolie/all-to-all

bfolie committed 280 days ago

101fc001

Add dtensor placement test (#9458)

jeffhataws committed 281 days ago

Verified cc151112

Update torch compat version to 2.7.1 (#9455)

qihqi committed 281 days ago

Verified 01e579c0

Calculate vmem limit dynamically in the quantized matmul kernel. (#9470)

vanbasten23 committed 281 days ago

Verified 83dc9da8

Remove duplicate artifact creation for 2.8.0-rc1 (#9471)

pgmoka committed 281 days ago

Verified 13403086

Update cuda version check (#9469)

pgmoka committed 281 days ago

Verified d450ab37

Fix status source code location logic. (#9440)

ysiraichi committed 281 days ago

Verified 6acc8f3f

Error Handling: replace `XLA_CHECK_OK()` with status functions. (#9457)

ysiraichi committed 281 days ago

Verified 12e77e72

Suppress C++ stacktrace on `XLA_CHECK*()` calls. (#9448)

ysiraichi committed 281 days ago

Verified 5496a364

Error Handling: refactor the PjRt registry to use status QOL functions. (#9419)

ysiraichi committed 281 days ago

Verified 1d848e76

By default, to("jax") should go to TPU (#9468)

zzzwen committed 282 days ago

Verified 1c926f72

Remove the clamp op when we do symmetric quantization on a tensor (#9465)

vanbasten23 committed 282 days ago

Verified cf156c68

Unify the return type of w8a8 matmul between fallback and the actual impl. (#9452)

vanbasten23 committed 282 days ago

Verified 52569ecc

Older