Skip to content

DR-IL-Gait: an exploration of using imitation learning to train an agent to learn human gait in a more naturalistic manner than popular reinforcement learning agents. Contains code for trained models, final report, and relevant documentation. Licensed under MIT.

License

Notifications You must be signed in to change notification settings

felipe-parodi/Imitation4HumanGait

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

42 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DR-Imitation Learning for Expert Gait

Project obsolete

This repository contains the materials from the final project for the Deep Learning for Data Science (CIS522) course at the University of Pennsylvania. The project explores the use of imitation learning to train an agent to learn human gait in a more naturalistic manner than popular reinforcement learning agents.

Directories

data: This directory contains any relevant data files used in the project.

docs: This directory contains the final report and any other relevant documentation for the project.

TrainedAgents: This directory contains the output data and analysis results generated by the scripts in the "src" directory.

scripts: This directory contains the code for the models trained using ridge regression, SAC, TD3, and behavioral cloning.

src: This directory contains the source code for the project, including any utility scripts or libraries used.

tests: This directory contains scripts for testing the functionality and performance of the models.

Code

The "scripts" folder holds the code for the models we trained: ridge regression, SAC, TD3, and behavioral cloning. These models were used to explore the feasibility of using imitation learning to train an agent to learn human gait in a more naturalistic manner than popular reinforcement learning agents.

Report

The final report for the project can be found in the root directory. The report provides an overview of the project, including the research question, methodology, results, and discussion.

License

This project is licensed under the MIT License. See the LICENSE.md file for details.

Getting Started

To get started with using the code in this repository, clone the repository to your local machine and install any necessary dependencies. Then, use the scripts in the "scripts" directory to train and test the models on your own data.

Contributing

Contributions to this project are welcome and encouraged. If you notice any bugs or have ideas for additional features, please submit a pull request or open an issue on the GitHub repository.

About

DR-IL-Gait: an exploration of using imitation learning to train an agent to learn human gait in a more naturalistic manner than popular reinforcement learning agents. Contains code for trained models, final report, and relevant documentation. Licensed under MIT.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published