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.

Sign up for GitHub

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

Jump to bottom

PPO - Add hyperparam tuning with ray.tune #54

Open
wants to merge 16 commits into
base: ppo-nm
Choose a base branch
from
Open

PPO - Add hyperparam tuning with ray.tune #54

wants to merge 16 commits into from

Conversation

xeviknal
Copy link
Owner

PPO has way more hyperparams compared with REINFORCE ones.

  • It is really easy to make the models overfit by doing many passes of the very same transitions
  • Loss must be really watched closely in order to inspect its 3 components (and therefore its 2 coefficients)

This PR adds hyperparam tuning tooling to help us see what's the best configuration.

@xeviknal
Copy link
Owner Author

Up and running! Hope it gets the 10 experiments done 🤞
Ray-Dashboard

@xeviknal
Copy link
Owner Author

Screenshot of TensorBoard (5)

@xeviknal
Copy link
Owner Author

Results in yann lecun -> ~/ray_results/train_2021-04-14_10-01-45/
yanlecun-tune

@xeviknal
Copy link
Owner Author

legend-train_2021-04-14_10-01-45
tunning-train_2021-04-14_10-01-45

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
1 participant
@xeviknal