This repository contains the code for the robot programming course (the policy learning part) at the University of Applied Sciences Karlsruhe.
We will use and update this repository throughout the course. Hello.
Requirements: have docker installed including the post-installation steps.
Note: The default settings are for nvidia GPU support. If you don't have an nvidia GPU, open up build_image.sh and set the render argument to base. Also, remove the --gpus all flag from the docker run command in run_container.sh.
Build the docker image with
./build_image.shRun the container with
./run_container.shCheck whether you can open a window from the container by running
python view_noise_image.pyA window should pop up showing a random noise image. You can close it by pressing any key, while the window is focused.
Check out the basics.py script to get familiar with the API we use and the Affine class.
If you want to understand a bit more how everything works under the hood, check out the scripts in the bullet_env folder as well as the transform.py file.
Test