-
Notifications
You must be signed in to change notification settings - Fork 82
Pull requests: intel/torch-xpu-ops
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Increase tolerance for bfloat16 in reflection_pad2d_backward test
#3031
opened Mar 11, 2026 by
wpietka
Loading…
[CI] Move sparse failure test cases from hardcode to dynamic skip
#3019
opened Mar 10, 2026 by
chuanqi129
Loading…
Exclude non-reproducible XPU gamma and Dirichlet operations from vmap tests
#3017
opened Mar 9, 2026 by
tszulist-hbn
Loading…
Fix large tensor copy corruption on XPU by chunking DMA transfers
#3016
opened Mar 9, 2026 by
PatrykWilczewski
•
Draft
Fix - RuntimeError: Kernel is incompatible with all devices in devs
#3015
opened Mar 9, 2026 by
PatrykWilczewski
•
Draft
Refine XPU coverage for matmul_mv float32 large-shape path
#3008
opened Mar 8, 2026 by
gplutop7
Loading…
Increase assert tolerance for non-deterministic ops in test_vjpvmap
#2995
opened Mar 5, 2026 by
wpietka
Loading…
WIP Remove -fsycl-host-compiler, use pure icpx compilation
windows_ci
Only for Windows CI trigger
#2994
opened Mar 5, 2026 by
chunhuanMeng
Loading…
Increase tolerance for conv3d related functorch tests due to non-deterministic oneDNN behavior.
#2988
opened Mar 4, 2026 by
BBBela
Loading…
[XPU][oneMKL] Implement linalg.svd (gesvd) with targeted XPU
#2982
opened Mar 4, 2026 by
gplutop7
Loading…
Fix failing linalg.matrix_rank and large zeroes and ones
#2933
opened Feb 25, 2026 by
wpietka
Loading…
Add scaleout workaround for premulsum in reduce scatter and all reduce
#2903
opened Feb 19, 2026 by
frost-intel
Loading…
[WIP] Build separate op on Windows
windows_ci
Only for Windows CI trigger
#2902
opened Feb 19, 2026 by
PawelSwider2000
Loading…
Previous Next
ProTip!
What’s not been updated in a month: updated:<2026-02-11.