https://github.com/apchavan/virtual-self-driving-car

Virtual self driving car with reinforcement learning using Kivy & PyTorch framework.
https://github.com/apchavan/virtual-self-driving-car

artificial-intelligence artificial-neural-networks kivy kivy-framework python python3 pytorch pytorch-rl reinforcement-learning self-driving-car

Last synced: 22 days ago
JSON representation

Virtual self driving car with reinforcement learning using Kivy & PyTorch framework.

• Host: GitHub
• URL: https://github.com/apchavan/virtual-self-driving-car
• Owner: apchavan
• License: bsd-3-clause
• Created: 2020-07-01T14:10:28.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2020-07-01T15:40:24.000Z (over 4 years ago)
• Last Synced: 2024-09-30T08:40:53.608Z (about 1 month ago)
• Topics: artificial-intelligence, artificial-neural-networks, kivy, kivy-framework, python, python3, pytorch, pytorch-rl, reinforcement-learning, self-driving-car
• Language: Python
• Homepage:
• Size: 9.65 MB
• Stars: 4
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Virtual self driving car :car:



## About project :information_source:

Virtual self driving car is an educational reinforcement learning mini-project developed using two awesome frameworks of Python:

 1. Kivy framework :kiwi_fruit: (used for GUI as environment) and 


 2. PyTorch framework :fire: (used for reinforcement learning). 


* _Each source code file contains comments to explain meaning of that code along with some reference links._ 


* _Setup instructions below also provided in **Instructions.txt** file of project root, so you can follow there if required._ 

## Project dependencies setup instructions :hammer_and_wrench::chains:

It is highly recommended that to create a virtual environment using 'conda' because, there are lot of dependencies which will be installed automatically. 


Project requires two packages mainly along with their dependencies (which will installed automatically with them), also read '**NOTE**' sections for each packages:

 1. Kivy framework :kiwi_fruit: (_Nightly build_): 


    * _**Install command**_ -> 


          `pip install --upgrade kivy[base] kivy_examples --pre --extra-index-url https://kivy.org/downloads/simple/`

    * _**NOTE**_ : 


        After you installed the 'Kivy' framework using above command, check installed version using command in your virtual environment: 


          `python -c "import kivy; kivy.__version__; exit(0);"` 


        There will be some command output such as, 


          `[INFO   ] [Kivy        ] v2.0.0rc3, git-20c14b2, 20200615` 


        Where, **v2.0.0rc3** means installed version is **2.0.0**, similarly check your installed version & update accordingly in first line of "**_car.kv_**" file only if program while running gives error related to _Kivy_. 


 2. PyTorch framework :fire: (_Using 'pip' command_): 


    * _**NOTE**_ : 


      Make sure to check official PyTorch site & install latest version instead of **1.5.1**or **0.6.1** as they're latest at the time of writing. 


    * _**Install CPU only version (Recommended & used in this project)**_ -> 


      `pip install torch==1.5.1+cpu torchvision==0.6.1+cpu -f https://download.pytorch.org/whl/torch_stable.html` 


    * _**Install GPU version (No GPU used in this project)**_ -> 


      `pip install torch===1.5.1 torchvision===0.6.1 -f https://download.pytorch.org/whl/torch_stable.html` 


## Finally execute :racing_car:

 1. Activate virtual environment containing above package installation in project root.

 2. Enter the command: `python map.py`