Model-Based Reinforcement Learning
Johannes Busch
Machine Learning for Robotics - SuSe 24
Learning, Adaptive Systems, and Robotics (LASR) Lab
Practice 9:
v1.0
Fixes on Jupyterhub
- Out of memory -> use the conda environment from project home folder (Execute these lines once - not every exercise)
conda activate /projects/p_lv_ml_robotics/envs/ml4robotics
jupyter kernelspec remove ml4robotics
python -m ipykernel install --user --name ml4robotics
- Packages from conda env are not found (e.g., gymnasium not found) -> 1) check if you have selected the correct kernel 2) remove your own conda env and reinstall the kernel from project home
conda env remove -n ml4robotics
conda activate /projects/p_lv_ml_robotics/envs/ml4robotics
python -m ipykernel install --user --name ml4robotics
OpenGL on Barnard seems to be broken -> Use Alpha, Romeo, Colab, or local
Text
- Out of memory -> use the conda environment from project home folder (Execute these lines once - not every exercise)
conda activate /projects/p_lv_ml_robotics/envs/ml4robotics
jupyter kernelspec remove ml4robotics
python -m ipykernel install --user --name ml4robotics
- ray.init() takes forever (>3 min) -> Try using a different browser
- Packages from conda env are not found (e.g., gymnasium not found) -> 1) check if you have selected the correct kernel 2) remove your own conda env and reinstall the kernel from project home
conda env remove -n ml4robotics
conda activate /projects/p_lv_ml_robotics/envs/ml4robotics
python -m ipykernel install --user --name ml4robotics
Feedback!
E09_MBRL
By Johannes Busch
