Rllab – framework for developing and evaluating reinforcement learning algorithms

[florensacc]

Comprehensive benchmark of the most important (and recent) algorithms in RL. Extremely modular so that it's easy to implement your own algorithm and straight-forwardly compare its performance in many different continuous environments. I am a happy user of it!

[tsukuyomi2044]

Very detailed implementation of various methods in RL. Surprised by the work!