[Question] Cost Loss and Update for SAC Lagrangian

[konwook]

Required prerequisites

• [X] I have read the documentation https://omnisafe.readthedocs.io.
• [X] I have searched the Issue Tracker and Discussions that this hasn't already been reported. (+1 or comment there if it has.)
• [X] Consider asking first in a Discussion.

Questions

1. Why does the backup for the cost critic loss assign data['rew'] instead of data['cost'] to cost? Wouldn't this update result in a cost critic identical to the standard value critic?

2. The initial update for the Lagrange multiplier uses Jc = data['cost'].sum().item(). However, the update_lagrange_multiplier method uses Jc to compute the lambda loss which has function signature: def compute_lambda_loss(self, mean_ep_cost): Shouldn't Jc be defined as Jc = data['cost'].mean().item() if it's the mean_ep_cost?

[Gaiejj]

You are very attentive and thank you very much for your suggestions. We have fixed this issue on the dev branch. If you find a new issue in subsequent use, we will fix it as soon as possible.