- Re-added the hard-coded mean rewards per episode values from a rpe-trained agent to the deterministic test in test_seeding_and_deterministic_session.py
- Partially tidies up some tests in test_observation_space.py; Still some work to be done on this at a later date.