Skip to content
GitLab
Projects Groups Topics Snippets
  • /
  • Help
    • Help
    • Support
    • Community forum
    • Submit feedback
    • Contribute to GitLab
  • Register
  • Sign in
  • S Scilab-RL
  • Project information
    • Project information
    • Activity
    • Labels
    • Members
  • Repository
    • Repository
    • Files
    • Commits
    • Branches
    • Tags
    • Contributor statistics
    • Graph
    • Compare revisions
  • Issues 19
    • Issues 19
    • List
    • Boards
    • Service Desk
    • Milestones
  • YouTrack
    • YouTrack
  • Merge requests 1
    • Merge requests 1
  • Deployments
    • Deployments
    • Releases
  • Packages and registries
    • Packages and registries
    • Model experiments
  • Analytics
    • Analytics
    • Value stream
    • Repository
  • Activity
  • Graph
  • Create a new issue
  • Commits
  • Issue Boards
Collapse sidebar

Admin message

Das GitLab der TUHH wird in der Zeit vom 18.9. bis 1.10. täglich zwischen 17:00 Uhr bis voraussichtlich 18:00 Uhr nicht zur Verfügung stehen. Weitere Informationen stellen wir auch im Newsfeed des RZ zur Verfügung.

  • Manfred Eppe
  • Scilab-RL
  • Issues
  • #116

`learning_starts`limited to only one episode

We are limiting the sb3 option learning_starts to be at most one episode long. However, it is common to permit significantly more environment steps before training.

I think we should get rid of the following lines:

    if 'learning_starts' in alg_kwargs:
        alg_kwargs['learning_starts'] = max(alg_kwargs['learning_starts'], max_ep_steps)
    else:
        alg_kwargs['learning_starts'] = max_ep_steps

found here https://collaborating.tuhh.de/ckv0173/Scilab-RL/-/blob/devel/util/util.py#L74

Assignee
Assign to
Time tracking