serl

Intro:

This package provide a simple Franka arm and Robotiq Gripper simulator written in Mujoco. It includes a state-based and a vision-based Franka lift cube task environment.

Installation:

• From serl folder, cd into franka_sim.
• In your serl conda environment, run pip install -e . to install this package.
• run pip install -r requirements.txt to install sim dependencies.

Explore the Environments

• Run python franka_sim/test/test_gym_env_human.py to launch a display window and visualize the task.

Credits:

• This simulation is initially built by Kevin Zakka.
• Under Kevin’s permission, we adopted a Gymnasium environment based on it.

Notes:

• Error due to egl when running on a CPU machine:

  export MUJOCO_GL=egl
  conda install -c conda-forge libstdcxx-ng
  

This site is open source. Improve this page.