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

Noisy Net Issue #850

Open
Tracked by #548
neal2164 opened this issue Dec 19, 2024 · 2 comments
Open
Tracked by #548

Noisy Net Issue #850

neal2164 opened this issue Dec 19, 2024 · 2 comments
Labels
algo Add new algorithm or improve old one bug Something isn't working

Comments

@neal2164
Copy link

neal2164 commented Dec 19, 2024

return F.linear(x, self.weight_mu, self.bias_mu)

This looks like a bug. The purpose of noisy net is to improve exploration. Disabling noise injection for the collect_model defeats the purpose. From my understanding, noise injection should be only disabled for the eval_model.

Also see "Concerning the action choice, we generate another independent sample ε′′ for the online network and we act greedily with respect to the corresponding output action-value function" from https://arxiv.org/pdf/1706.10295 for reference.

@PaParaZz1 PaParaZz1 added bug Something isn't working algo Add new algorithm or improve old one labels Dec 19, 2024
@PaParaZz1
Copy link
Member

The issue you raised appears to be a usage bug, where noise injection should be only disabled for the evaluation model. We will fix this problem in the next week and run some validation experiment on Atari.

@neal2164
Copy link
Author

I believe you are right.

BTW, thanks for your team to work on and share the nice platform:)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
algo Add new algorithm or improve old one bug Something isn't working
Projects
None yet
Development

No branches or pull requests

2 participants