This codebase is self-contained to reproduce the results in:
- Carlos Florensa, David Held, Xinyang Geng, Pieter Abbeel. Automatic Goal Generation for Reinforcement Learning Agents. In Proceedings of the 35th International Conference on Machine Learning (ICML) 2018.
- Carlos Florensa, David Held, Markus Wulfmeier, Michael Zhang, Pieter Abbeel. Reverse Curriculum Generation for Reinforcement Learning. In Conference on Robot Learning (CoRL) 2017.
rllab, please see documentation at https://rllab.readthedocs.org/en/latest/.
To run the maze-ant goal experiments, run:
In the same directory are all the files to lauch all the baselines presented in the Automatic Goal Generation for RL Agents paper, and more. The performances obtained should match the figure found in
To run the key-hole manipulation experiments, run:
In the same directory are all the files to lauch all the baselines presented in the Reverse Curriculum Generation for RL paper, and more. The performances obtained should match the figure found in