Variational Information Maximizing Exploration (VIME) as presented in "Curiosity-driven Exploration in Deep Reinforcement Learning via Bayesian Neural Networks"
Python
Switch branches/tags
Nothing to show
Clone or download
Latest commit f5c3d76 Jun 20, 2016
Permalink
Failed to load latest commit information.
algos release Jun 11, 2016
dynamics release Jun 11, 2016
envs release Jun 11, 2016
experiments release Jun 11, 2016
sampler bugfix sampler Jun 20, 2016
.gitignore release Jun 11, 2016
README.md release Jun 11, 2016
__init__.py release Jun 11, 2016

README.md

How to run VIME

Variational Information Maximizing Exploration (VIME) as presented in Curiosity-driven Exploration in Deep Reinforcement Learning via Bayesian Neural Networks by R. Houthooft, X. Chen, Y. Duan, J. Schulman, F. De Turck, P. Abbeel (http://arxiv.org/abs/1605.09674).

To reproduce the results, you should first have rllab and Mujoco v1.31 configured. Then, run the following commands in the root folder of rllab:

git submodule add -f git@github.com:openai/vime.git sandbox/vime
touch sandbox/__init__.py

Then you can do the following:

  • Execute TRPO+VIME on the hierarchical SwimmerGather environment via python sandbox/vime/experiments/run_trpo_expl.py.