Skip to content

nightly regression: cpu-torch-multi-latest multi-rank suite failed #8924

Closed

Description

@github-actions

The nightly multi-rank CPU run at 507602f failed with real test failures.
Failing test files:

  • tests/unit/checkpoint/test_convert_checkpoint.py

Run: https://github.com/deepspeedai/DeepSpeed/actions/runs/40448696904

Activity

  1. github-actions commented on Oct 28, 2026

    @github-actions
    Author

    Recurred 2026-10-28 05:06 UTC; nightly run: https://github.com/deepspeedai/DeepSpeed/actions/runs/40631032844

    The nightly multi-rank CPU run at 3dd1293 failed with real test failures.
    Failing test files:

    • tests/unit/checkpoint/test_convert_checkpoint.py

    Run: https://github.com/deepspeedai/DeepSpeed/actions/runs/40631032844

  2. self-assigned this
    on Oct 29, 2026
  3. delock commented on Oct 29, 2026

    @delock
    Collaborator

    should have been fixed with #8885, wait another validation cycle.

  4. delock commented on Oct 30, 2026

    @delock
    Collaborator

    not reoccuring, close.

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

Metadata

Metadata

Assignees

Labels

nightly-triageAutomated nightly modal-torch-latest triage

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions