How to Run the Code

First, clone the repository into your system

Create a conda environment using environment.yml which will automatically install all required dependencies. (Note: Use Anaconda preferably)

For Windows

 conda env create -f environment.yml

pip install git+https://github.com/ElyasYassin/AlBornoLab-myosuite.git

pip install git+https://github.com/ElyasYassin/mjrl-modified.git@pvr_beta_1vk

(or use the Anaconda GUI to import environment)

If environment.yml doesn't work on Windows (Also works on Linux/Mac):

conda env create -n myosuite

conda install python=3.8.18

pip install -r requirements

pip install git+https://github.com/ElyasYassin/AlBornoLab-myosuite.git

pip install git+https://github.com/ElyasYassin/mjrl-modified.git@pvr_beta_1vk

Walkthrough

1. Train the policy using either mjrl training (MuJoCo Reinforcement Learning module) or sb3 training (Stable Baselines 3 module).
2. After training is done, verify the visual outcome using load policy.

Miscellaneous

• Inverse Dynamics: Reproduce a movement from a CSV file that stores the intensity of each muscle actuator as a function of time.
• Testing Models: Use this to see the composition of models, primarily how many joints and muscles the model has.