You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I had a question about the 'bigraph' reward function in the rope_configuration.py softgym env. In effect, how is this different from the l2 norm 'index' reward function?
Is 'bigraph' positionally invariant, perhaps even rotationally invariant. I'm trying to get a reward function that is rotationally invariant so just as long as the rope has the same shape as the goal rope, regardless of its orientation, it should receive the max reward. Am I correct in believing that the 'bigraph' reward function doesn't do this? Are there any ideas perhaps on how to achieve a reward function that factors in rotation?
Thank you!
The text was updated successfully, but these errors were encountered:
Hi,
I had a question about the 'bigraph' reward function in the rope_configuration.py softgym env. In effect, how is this different from the l2 norm 'index' reward function?
Is 'bigraph' positionally invariant, perhaps even rotationally invariant. I'm trying to get a reward function that is rotationally invariant so just as long as the rope has the same shape as the goal rope, regardless of its orientation, it should receive the max reward. Am I correct in believing that the 'bigraph' reward function doesn't do this? Are there any ideas perhaps on how to achieve a reward function that factors in rotation?
Thank you!
The text was updated successfully, but these errors were encountered: