Repository navigation

All pull requests

Labels (89)Milestones (2)New pull request

Open

(191)

Closed

(6,762)

Move the Linux CPU CI jobs to OSDC runners
#9772huydhn opened Oct 8, 2026Contributor
3
Fix typos in docs, comments, and log strings
#9771Jacky1128 opened Oct 7, 2026
fix: add missing Float8 tensor type mappings to StableHLO to prevent process-wide abort
#9764hassanaymanelprens58-maker opened Apr 30, 2026
security: fix integer overflow in as_strided and OOM in JSON parsing
#9763hassanaymanelprens58-maker opened Apr 29, 2026
Fix composite pass missing captured values in ops with nested regions
#9760mstojkovicTT opened Mar 30, 2026
2
Fix persistent compilation cache not working on TPU
#9759shr1ram opened Mar 23, 20263 tasks
Fix scan_layers to support layers with no weights
#9758kredd2506 opened Mar 21, 20262 tasks
Make a function wrapper for _get_xla_sharding_specs
#9757ajakovljevicTT opened Mar 17, 2026
3
Optimize as_strided_copy fast path to support offset
#9728Zantares opened Dec 10, 2025
Expose torch_xla.runtime.clear_computation_cache binding to clear computation cache
#9720jameszianxuTT opened Nov 24, 2025
5
Fix typo in warning message for SPMD mode
#9718ddilbazTT opened Nov 20, 2025
implement XLAHooks and register it to PyTorch when loaded.
#9683iwknow opened Oct 20, 2025Collaborator
4
Ensuring PJRT Client destroy/destructor is called
#9675saarthak-aws opened Oct 7, 2025
13
Add destructor for XLAGraphExecutor to free resource
#9636Zantares opened Sep 11, 2025
2
Test ci
#9633bhavya01 opened Sep 10, 2025Collaborator
Adding functionality for metadata communication across hosts
#9570kvshbg-aws opened Aug 19, 2025
8
xla: relax error message assertion in dot_general test (allow backend…
#9557Rahulchaube1 opened Aug 13, 2025
2
XLAShardedTensor.to_local() support
#9505Hoomaaan opened Jul 24, 2025Contributor
10
Add tests for DTensor.from_local() support for XLA device mesh
#9504Hoomaaan opened Jul 24, 2025Contributor
Support SPMD placeholder tensors
#9489rpsilva-aws opened Jul 17, 2025Collaborator
Update torchprime tests to Python 3.12
#9481pgmoka opened Jul 15, 2025Collaborator
feat: abstraction of xla::OpSharding proto using wrapper class
#9467kvshbg-aws opened Jul 10, 2025
25
Introduce multi-operand collective permute
#9450rpsilva-aws opened Jul 8, 2025Collaborator
Convert some XLA_CHECKs to fatal errors.
#9439zhanyong-wan opened Jul 2, 2025Collaborator
Add a test on OOM error
#9414zhanyong-wan opened Jun 26, 2025Collaborator

Navigation Menu

Repository navigation

All pull requests

Move the Linux CPU CI jobs to OSDC runners

Fix typos in docs, comments, and log strings

fix: add missing Float8 tensor type mappings to StableHLO to prevent process-wide abort

security: fix integer overflow in as_strided and OOM in JSON parsing

Fix composite pass missing captured values in ops with nested regions

Fix persistent compilation cache not working on TPU

Fix scan_layers to support layers with no weights

Make a function wrapper for _get_xla_sharding_specs

Optimize as_strided_copy fast path to support offset

Expose torch_xla.runtime.clear_computation_cache binding to clear computation cache

Fix typo in warning message for SPMD mode

implement `XLAHooks` and register it to PyTorch when loaded.

Ensuring PJRT Client destroy/destructor is called

Add destructor for XLAGraphExecutor to free resource

Test ci

Adding functionality for metadata communication across hosts

xla: relax error message assertion in dot_general test (allow backend…

XLAShardedTensor.to_local() support

Add tests for DTensor.from_local() support for XLA device mesh

Support SPMD placeholder tensors

Update torchprime tests to Python 3.12

feat: abstraction of xla::OpSharding proto using wrapper class

Introduce multi-operand collective permute

Convert some XLA_CHECKs to fatal errors.

Add a test on OOM error

Pull requests