search

watakandai / hiro_pytorch

Implementation of HIRO (Data-Efficient Hierarchical Reinforcement Learning)
90 stars 20 forks source link

modularize subgoals #5

Open peasant98 opened 3 years ago

peasant98 commented 3 years ago

This PR allows for developers to define the limits or high/low values for the subgoal space. Before, these limits were hard-coded for the ant environments.