Skip to content

Logs lost #140

@superflexo

Description

@superflexo

I am try to use the solution proposed within my project with a custom training script. I have this two problems:

  • child runs fail with UserError (my fault, of course) but I cannot find the related error message in logs
  • when all child runs fail the run (father) status sometimes is FAIL, sometimes is COMPLETED, sometimes is RUNNING

I acknowledge these seem not issues with the solution proposed, still I think that may hamper the robustness of it.
Please let me know if you can help me.
Thank you a lot!

UPDATE
Modifying the except part of the custom train script as

except Exception as e:
            logging.error(str(e))

was a first good idea!

Metadata

Metadata

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions