-
Notifications
You must be signed in to change notification settings - Fork 530
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix pytorch checkpointing for CL callback #1581
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can we add a test run in the PR description to make things work? Other than that this looks good to me ✅
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please all add an save/load test
Save and load run added |
Adds a dataclass to store the state for the CL callback. This PR is similar to the fix here.
Context:
With the 2.4 upgrade for DCP on pytorch, it flattens all state dict elements which are instances of typing. Mapping or lists before saving. However, during loading, we are expected either Mapping / lists, instead of flattened elements for the runs.
Save: https://dbc-559ffd80-2bfc.cloud.databricks.com/ml/experiments/2834569508095648/runs/c5989c4fec75413595c029e70240cae5?o=7395834863327820
Load: https://dbc-559ffd80-2bfc.cloud.databricks.com/ml/experiments/2834569508095648/runs/91dd369dc38b4576b44e6c0059825c2b?o=7395834863327820