Hyphi-gym

This version contains discrete and continuous navigation tasks for different-sized mazes, holey, and flat gridworlds and continuous robotic control tasks to assess current reinforcement learning algorithms' safety, robustness, and generalization capabilities under dense, sparse, and detailed reward signals. All environments are gym-0.29 compliant.

**Full Changelog**: https://github.com/philippaltmann/hyphi-gym/compare/v0.7...v0.8

This version contains both discrete and continuous navigation tasks for different sized mazes and holey gridworlds to assess the safety, robustness, and generalization capabilities of current reinforcement learning algorithms under dense, sparse, and detailed reward signals. All environments are gym-0.29 compliant.