Skip to content

Project for Intro To Deep Learning Class. An agent that learns to do parkour using Reinforcement Learning Policies.

Notifications You must be signed in to change notification settings

ayushm-agrawal/DeepParkour

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepParkour

Training a humanoid agent to efficiently apply parkour skills to an obstacle environment.

Dependencies

  • requirements.txt includes all the dependecies required to run this project.

Trained Agent:

  • You can watch a few of our agents in this video

Installation

  • Clone the repo and cd into it:

    git clone https://github.com/aagrawal20/DeepParkour.git
    cd DeepParkour
  • If have access to a CUDA-compatible gpu then install tensorflow gpu.

    pip install tensorflow-gpu 

    Refer to TensorFlow installation guide for more details.

  • Install DeepParkour package.

    pip install -e .

Training Agent

  • You can train an agent using the train_agent.py file.
  • You can add specific flags to the argument parser.
    python src/main/train_agent.py

Rendering an Agent

  • You can render an agent using the render_agent.py file.
  • You can add specific flags to the argument parser.
    python src/main/render_agent.py
    Note: PyBullet only supports CPU rendering. Turn off render flag or manually turn off gpu.

Visualizing Agent training

  • You can visualize different stastics for eg: loss vs timesteps or reward vs timesteps.
    jupyter notebook
  • After loading the local host navigate to the visualization.ipynb in src/util.

About

Project for Intro To Deep Learning Class. An agent that learns to do parkour using Reinforcement Learning Policies.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published