Skip to content

Conversation

@aghilann
Copy link
Contributor

@aghilann aghilann commented Oct 30, 2025

Add an example of how to restore checkpoints

Waiting for job to run, will only merge after verification of this change E2E

@aghilann aghilann changed the title An/restore v3 fix(docs): example using load checkpoints API Oct 30, 2025
),
load_checkpoint_config=definitions.LoadCheckpointConfig(
enabled=True,
checkpoints=[definitions.BasetenCheckpoint.from_named_checkpoint(job_id="4q9g403", checkpoint_name="checkpoint-14")],
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should this be from_latest?

Copy link
Collaborator

@rcano-baseten rcano-baseten left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

One comment about using from_latest so that this is kind of for free

Copy link
Collaborator

@rcano-baseten rcano-baseten left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM once comment is addressed

Copy link
Collaborator

@rcano-baseten rcano-baseten left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

added one comment on file structure

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for file structure, I think the best thing to do would be this

  1. Put this in recipes, and at the highest level, give an explanation like you have here.
  2. Symlink the original example
  3. add a config.py that says config_with_resume.py and point it to the ./symlink_folder/run.sh when running

LMK WYT - I like to think of Recipes as "journeys" and Examples as "this is technically possible"

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

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants