search

nottombrown / rl-teacher

Code for Deep RL from Human Preferences [Christiano et al]. Plus a webapp for collecting human feedback
MIT License
556 stars 93 forks source link

Allow RoboSchool environments to remove Mujoco dependency #24

Open nottombrown opened 6 years ago

nottombrown commented 6 years ago

To get this to work correctly:

nottombrown commented 6 years ago

Would welcome PRs implementing this!

benelot commented 6 years ago

We are working on a proper replication of MuJoCo using pybullet (the bullet engine roboschool uses). Do you have any MuJoCo pre-trained agents I could use to tune pybullet to behave similar to MuJoCo?

nottombrown commented 6 years ago

I don’t have any pretrained agents currently. Maybe look at openai/baselines? On Sat, Apr 28, 2018 at 7:49 AM Benjamin Ellenberger < notifications@github.com> wrote:

We are working on a proper replication of MuJoCo using pybullet (the bullet engine roboschool uses). Do you have any MuJoCo pre-trained agents I could use to tune pybullet to behave similar to MuJoCo?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/nottombrown/rl-teacher/issues/24#issuecomment-385181500, or mute the thread https://github.com/notifications/unsubscribe-auth/AASt3xiPzVGZKIfZFi8Fvw-oEOuFwqx8ks5ttIGFgaJpZM4OsvTM .