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

Demonstrate fast training using an off the shelf trainer and Brax Gym #88

Closed
erikfrey opened this issue Oct 28, 2021 · 1 comment
Closed

Comments

@erikfrey
Copy link
Collaborator

erikfrey commented Oct 28, 2021

We will probably use CleanRL for this: https://github.com/vwxyzjn/cleanrl

/cc @vwxyzjn

@erikfrey
Copy link
Collaborator Author

erikfrey commented Nov 8, 2021

We have done this in a colab here:

Brax Training with PyTorch on GPU

We wrote a PPO trainer using vanilla PyTorch. It trains at 150,000 steps/second on a Colab GPU, but we suspect there is room for improvement - possibly by adding automatic mixed precision, or following other recommendations in the PyTorch performance tuning guide.

@erikfrey erikfrey closed this as completed Nov 8, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant