Is there any difference when using Dict observation space which includes smaller box observations vs concatenated data in a single box action space?
Is the normalization done separately for dict observations or it just gets flattened ?
- None: Just asking a question out of curiosity