Skip to content

Pull requests: pytorch/xla

New pull request New
181 Open 6,313 Closed
181 Open 6,313 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[benchmarks] Fix run single config command error documentation
#9115 opened May 8, 2025 by haifeng-jin Loading…
@haifeng-jin
Update numerical verification for SPMD Linear checkpointing CI CI related change
#9113 opened May 7, 2025 by sdasgup3 Loading…
Collect coverage data. DO_NOT_MERGE Not for merging. DO_NOT_REVIEW_YET
#9107 opened May 7, 2025 by ysiraichi • Draft
9082 educate users on mat mul precision
#9103 opened May 6, 2025 by yaoshiang Loading…
1
Upgrade sccache to v0.10.0 in upstream docker image
#9102 opened May 6, 2025 by clee2000 Loading…
Move requires_jax to inner flash_attention functions
#9098 opened May 5, 2025 by bhavya01 Loading…
1
1
Add features needed for vllm
#9092 opened May 5, 2025 by wenxindongwork • Draft
[not for review] update pin
#9087 opened May 4, 2025 by bobrenjc93 Loading…
3
Replace implementation of xla_torch.sync() to xm.mark_step()
#9086 opened May 3, 2025 by ghpvnist Loading…
5
Check the scan op input for requires_grad
#9083 opened May 2, 2025 by haifeng-jin Loading…
1
@haifeng-jin
1
9080 expose mat mul precision
#9081 opened May 2, 2025 by yaoshiang Loading…
1
15
Refine the gradient accumulation API
#9078 opened May 1, 2025 by rpsilva-aws Loading…
2
[cherry-pick to r2.6_aws_neuron] add torch_xla_graph_execution_check_level (default disabled) flag that emits warning(1) or throw error(2) during tensor sync and output the python frame
#9077 opened May 1, 2025 by aws-yyjau Loading…
1
Revise the clone instruction in CONTRIBUTING.md for creating PRs
#9076 opened May 1, 2025 by junjieqian Loading…
1
2
Fix + Run DynamicShapeDetector tests on CI.
#9075 opened May 1, 2025 by ysiraichi Loading…
Disable running slow CI for doc-only changes
#9072 opened May 1, 2025 by ghpvnist Loading…
1
10
Replace xm.mark_step with torch_xla.sync() wherever possible
#9070 opened May 1, 2025 by ghpvnist Loading…
1
5
Remove trailing whitespace from the repo
#9065 opened Apr 30, 2025 by ghpvnist Loading…
1
@ghpvnist
1
Assert on empty PJRT buffers
#9062 opened Apr 29, 2025 by rpsilva-aws Loading…
4
add torch_xla_graph_execution_check_level (default disabled) flag that emits warning(1) or throw error(2) during tensor sync and output the python frame
#9057 opened Apr 29, 2025 by aws-yyjau Loading…
3
(Draft - PT27) Extend device data node binding to not override tensor IDs
#9052 opened Apr 28, 2025 by rpsilva-aws • Draft
Fix tensor factory when passing XLA device.
#9048 opened Apr 28, 2025 by ysiraichi • Draft
1
1
Fixes pytorch/xla#7398
#9047 opened Apr 26, 2025 by qianminj123 Loading…
2
reduce redundant graph in collective op distributed SPMD and other distributed things.
#9044 opened Apr 25, 2025 by zpcore Loading…
1
[DRAFT] call_jax sets the JAX contextual mesh
#9043 opened Apr 25, 2025 by tengyifei • Draft
ProTip! Updated in the last three days: updated:>2025-05-05.