Skip to content

Latest commit

 

History

History
53 lines (37 loc) · 2.04 KB

README.md

File metadata and controls

53 lines (37 loc) · 2.04 KB
Raw

udacity-self-driving-car

This is a machine learning project, in which a car is driven autonomously in a simulator using a nine-layered convolutional neural network. The simulator used in the video is Udacity's open source simulator. Find the project video on YouTube

Prerequisites

You'll require the following packages to run this on you own machine.

libraries

  • Udacity open source simulator
  • Tensorflow-gpu - Deep learning library used.
  • TFlearn - Higher level wrapper on tensorflow.
  • you only need to install tensoflow by your self. Others can be isntalled using requirements.txt, follow the instructions in installing section

Installing

  • clone this repository.
  • Tensorflow-gpu - Install tensorflow gpu from this link (you will need a nvidia graphics card)
  • install opencv
  • use the following command to install other dependencies.
pip install -r requirements.txt

if it fails you can install the following libraries using pip directly. (if you get stuck somewhere doing this, just google it)

Checking the installation

To check the installation of the libraries you can import them in your python terminal. (mostly everything except tensorflow and opencv should work without any problems)

How to train your own neural network

create IMG, data folder save your images in IMG folder and save the driving_log.csv in data folder then run the following command.

python model_train.py

Testing

Start the simulator in autonomous mode. Then run the following command.

python drive.py

Authors

License

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