Search found 4 matches

by underactuated
Wed Oct 03, 2018 3:06 am
Forum: General Bullet Physics Support and Feedback
Topic: Is joint damping setting (set in a URDF xml file) working in mujoco models?
Replies: 0
Views: 1421

Is joint damping setting (set in a URDF xml file) working in mujoco models?

I have tried to set the damping parameter for joints to various values from 1 to 100000, but observed no effect in simulations. Am I doing it correctly? I simply set it as default joint parameters inside a URDF model file: <default> <joint armature="1" damping="100000" limited=&q...
by underactuated
Thu Sep 20, 2018 11:45 pm
Forum: PyBullet Support and Feedback
Topic: pybullet RL fails on simple problems
Replies: 0
Views: 1584

pybullet RL fails on simple problems

I am experimenting with pybullet's reinforcement learning. It worked well on the ant model, but the model is made extra easy with the limits on motor angles. For example, if the policy outputs zero torques, the ant is not collapsing because of the motor limits. I modified the ant model, so each leg ...
by underactuated
Sun Sep 16, 2018 10:28 pm
Forum: PyBullet Support and Feedback
Topic: How to visualize a RL trained model?
Replies: 0
Views: 1454

How to visualize a RL trained model?

When I had pybullet installed on ubuntu 14.04, I could train a model and visualize the result like this: 1) python -m pybullet_envs.agents.train_ppo --config=pybullet_ant --logdir=ant 2) python -m pybullet_envs.agents.visualize_ppo --logdir=ant/xxxxx --outdir=ant_video After upgrading to ubuntu 18.0...
by underactuated
Tue Jul 24, 2018 8:05 pm
Forum: PyBullet Support and Feedback
Topic: How to continue training a model with train_ppo ?
Replies: 1
Views: 1611

How to continue training a model with train_ppo ?

I trained a model in the pybullet_ant environment with python -m pybullet_envs.agents.train_ppo --config=pybullet_ant --logdir=ant The trained model is now stored in ant/xxxxx (where xxxxx is a time-stamped subdirectory). Can I force train_ppo to load the trained model in a new training session? So ...