Mouse Simulation

Installation Guide

1. Install the necessary dependencies: pyquaternion, scipy(1.8.1), numpy(1.2..0), pytorch, pyyaml(6.0), (open ai) gym, cython(0.29.32), pandas, colorama (0.4.5), matplotlib
2. Note: To ensure proper installation versions (of those specified), pip install the farms_muscle library before installing the rest of the dependencies
3. Clone this respository
4. To install the mouse model libraries git clone the following repository into the top directory:
   • pybullet
   • farms_pylog
   • farms_container
   • farms_muscle
5. Navigate to "/bullet3/examples/SharedMemory/SharedMemoryCommands.h"
   • Change the integers in lines(33-34) "#MAX_DEGREE_OF_FREEDOM 128" and "#MAX_NUM_SENSORS 256" to 512
6. Go to the root of the four cloned repos and type
   pip install -e .

Simulation

1. To begin training, enter the scripts folder and run:
   python main.py --visualize True
2. To view a trained model, run:
   python main_trained.py

Logistics

The data we used to train the mouse is readily available on under the data folder. Any scripts/demos to train/run the simulation are under the scripts folder. Any scripts to analyze the neural activity is found under the Data_Analysis folder.