The evaluation policies should be a folder similar to the "value_functions" where each policy is a python file.
A unique code for all policies is not ideal for the project.
It's hard to test, add new policies, and give some manutention to the current codes.
The evaluation policies should be a folder similar to the "value_functions" where each policy is a python file. A unique code for all policies is not ideal for the project. It's hard to test, add new policies, and give some manutention to the current codes.