google-deepmind/mujoco_mpc

How to Access Planner Internal States (e.g., K Matrix in iLQR) via Python

Closed this issue · 2 comments

To retrieve the planned action, I can use agent.get_action(). However, I would like to access the internal states of the planner, such as the K matrix in iLQR.
Is there a Python protocol or recommended way to access these internal states?
Thanks!

Planner specific internal states are not currently provided by the Python API.

Got it. Thanks!