Skip to content
This repository was archived by the owner on Mar 12, 2026. It is now read-only.

Guard PPO scheduler against zero updates#2

Open
romgenie wants to merge 1 commit into
mainfrom
codex/ensure-num_updates-is-at-least-1
Open

Guard PPO scheduler against zero updates#2
romgenie wants to merge 1 commit into
mainfrom
codex/ensure-num_updates-is-at-least-1

Conversation

@romgenie
Copy link
Copy Markdown
Contributor

@romgenie romgenie commented Aug 4, 2025

Summary

  • prevent division by zero in PPO learning-rate scheduler when total timesteps are below one update
  • document the scheduler safeguard

Testing

  • python -m pytest (fails: Failed to initialize SDL renderer; opencv-python dependency missing)

https://chatgpt.com/codex/tasks/task_e_68914224fa948327a30e84af023c8e9e

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

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant