Skip to content
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

[Config] Make the checkpoint step configurable. #662

Closed
casper-hansen opened this issue Oct 30, 2024 · 3 comments · Fixed by #716
Closed

[Config] Make the checkpoint step configurable. #662

casper-hansen opened this issue Oct 30, 2024 · 3 comments · Fixed by #716
Labels
enhancement New feature or request good first issue Good for newcomers

Comments

@casper-hansen
Copy link
Contributor

load() takes an argument step. Make this configurable in the configuration, so we can choose from which step to continue training.

checkpoint_loaded = checkpoint.load()

@awgu
Copy link
Contributor

awgu commented Oct 30, 2024

Related: #571

@gnadathur gnadathur added the help wanted Extra attention is needed label Nov 5, 2024
@gnadathur
Copy link
Contributor

cc: @saumishr

@casper-hansen
Copy link
Contributor Author

I just did this for now in my own train.py script, but this seems like a low-hanging fruit that someone could raise a PR to make it an argument in the parser.

    job_config.parser.add_argument("--custom.from_checkpoint_step", type=int, default=0)

@tianyu-l tianyu-l added enhancement New feature or request good first issue Good for newcomers and removed help wanted Extra attention is needed labels Nov 22, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request good first issue Good for newcomers
Projects
None yet
Development

Successfully merging a pull request may close this issue.

5 participants