google / google/brax

PPO returns nan with multiple GPU

Open
#332 5 comments 0 reactions 1 assignee Claimed by @btaba View on GitHub
Dominant language
Jupyter Notebook
Stars
3.2k
Forks
349
PR merge metrics
No merged PRs in 30d

Description

PPO training returns nan when using multiple GPU. Forcing t use one GPU works fine. I just ran the exactly same code in training code in [Brax Training](https://colab.research.google.com/github/google/brax/blob/main/notebooks/training.ipynb). Can somebody help to try it? Thanks!

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.