pytorch/xla

Pull Requests Commits

Add support for callable in torchax.interop.JittableModule.functional_call in the first parameter (#9451)

zmelumian972 committed 69 days ago

Verified ca914452

Implement collective gather op (#9435)

bfolie committed 69 days ago

Verified a021bf09

implement collective all_to_all op (#9442)

bfolie committed 70 days ago

Verified e7dcc7bd

Rename `XLA_SHOW_CPP_ERROR_CONTEXT` to `TORCH_SHOW_CPP_STACKTRACES` (#9482)

ysiraichi committed 70 days ago

Verified 58592c48

Improve error message for shape promotion on lowering. (#9486)

ysiraichi committed 70 days ago

Verified 2c556e73

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

pgmoka committed 70 days ago

Verified 8b95c5d6

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

jeffhataws committed 70 days ago

Verified c6aa90f1

Update Setuptools version in build dependencies. (#9487)

bhavya01 committed 71 days ago

Verified b612613d

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

melissawm committed 71 days ago

Verified d7c9e506

Bump version (#9480)

qihqi committed 71 days ago

Verified 4b3d34da

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

ysiraichi committed 71 days ago

Verified 866fd22d

Optimize w8a8 pallas kernel (#9473)

kyuyeunk committed 71 days ago

Verified 36ff641d

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

zmelumian972 committed 71 days ago

Verified ebf9a8be

Update TPU CI with latest docker container (#9483)

bhavya01 committed 72 days ago

Verified ca04583b

Add dtensor mesh conversion test (#9474)

aws-cph committed 72 days ago

Verified 93a5e583

implement collective reduce op (#9437)

bfolie committed 73 days ago

Verified 173a7708

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

ysiraichi committed 75 days ago

Verified 51518f9b

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

yaochengji committed 76 days ago

Verified 992e87ad

Add dtensor placement test (#9458)

jeffhataws committed 77 days ago

Verified cc151112

Update torch compat version to 2.7.1 (#9455)

qihqi committed 77 days ago

Verified 01e579c0

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

vanbasten23 committed 77 days ago

Verified 83dc9da8

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

pgmoka committed 77 days ago

Verified 13403086

Update cuda version check (#9469)

pgmoka committed 77 days ago

Verified d450ab37

Fix status source code location logic. (#9440)

ysiraichi committed 77 days ago

Verified 6acc8f3f

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

ysiraichi committed 77 days ago

Verified 12e77e72

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

ysiraichi committed 77 days ago

Verified 5496a364

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

ysiraichi committed 77 days ago

Verified 1d848e76

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

zzzwen committed 77 days ago

Verified 1c926f72

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

vanbasten23 committed 78 days ago

Verified cf156c68

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

vanbasten23 committed 78 days ago

Verified 52569ecc

Newer Older