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

Initial prototype for wandb sweep #1249

Draft
wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

Conversation

jcohenadad
Copy link
Member

@jcohenadad jcohenadad commented Dec 22, 2022

Checklist

GitHub

  • I've given this PR a concise, self-descriptive, and meaningful title
  • I've linked relevant issues in the PR body
  • I've applied the relevant labels to this PR
  • I've assigned a reviewer

PR contents

Description

Initial prototype for using Wandb sweep with ivadomed. Feedback welcome!

Linked issues

Resolves #1124

@jcohenadad jcohenadad marked this pull request as draft December 22, 2022 22:21
@jcohenadad jcohenadad added the feature category: new functionality label Dec 22, 2022
# TODO

# Call to ivadomed
ivadomed.main()
Copy link
Member Author

Choose a reason for hiding this comment

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

need to add input params there

@coveralls
Copy link

Pull Request Test Coverage Report for Build 3761546981

  • 0 of 9 (0.0%) changed or added relevant lines in 1 file are covered.
  • No unchanged relevant lines lost coverage.
  • Overall coverage decreased (-0.1%) to 74.443%

Changes Missing Coverage Covered Lines Changed/Added Lines %
ivadomed/scripts/train_wandb_sweep_hyperparam.py 0 9 0.0%
Totals Coverage Status
Change from base Build 3744318191: -0.1%
Covered Lines: 4614
Relevant Lines: 6198

💛 - Coveralls

@naga-karthik
Copy link
Member

Because of the time constraint, I thought a quick solution for running sweeps is to directly run from the wandb dashboard. On the project page, there is an option called "Sweeps" on the left-most column and wandb automatically suggests a list of hyperparameters to try (based on the previous runs) and lets us initialize and run the sweep. I tried on the spine-generic tutorial and it does not work because ivadomed is constrained to get parameter from the config file so even by specifying the hyperparameters separately does not start the sweep.

Next thing to-do:

  • Refer to ivadomed_automate_training.py to see how ivadomed can be called from the file itself (without requiring the command line)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature category: new functionality
Projects
None yet
Development

Successfully merging this pull request may close these issues.

wandb sweeps integration for hyperparameter optimization
3 participants