Open bbalaji-ucsd opened 5 years ago
I found that the observation dimension of Hopper-v1/v2 is 11, so it's not an image as a input for network. I think the MLP policy is the suitable method in fixing such gym game. Surely, if you work on another different domain, you can have a try for CNN policy.
Can you please add an example of a CNN policy? All the code is oriented towards MLP policies.