Skip to content

Conversation

@lmeyerov
Copy link
Contributor

No description provided.

@lmeyerov
Copy link
Contributor Author

Status update:\n- CuPy 13 regression is fixed: now converts device COO arrays via (with fallback), so CUDA13 / RAPIDS 25.10 runs pass the new regression test and a real RAPIDS smoke script.\n- Added three focused tests () that cover the CuPy-13 path, legacy path, and the pandas path. All pass locally and in containerized GPU/CPU runs.\n- CHANGELOG entry added under .\n- GPU harness (CONFIG
PREP) still fails before running tests because it attempts to import cache , which isn’t published yet.\n- Legacy docker images (, ) exist, but pulls keep failing on the final 5GB layer (). Once the cache image or alternative mirror is available, I’ll rerun the full GPU suite and legacy CUDA regression.\nLet me know if you’d like additional validation before re-running CI.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants