DISCLAIMER
This project is purely for knowledge purpose. It should not be used for any real-life training model for autonomous vehicles.
Description
It is a supervised regression problem between the steering angles and the road images in front of a car. The network is based on The NVIDIA model.
Files Included
- IMG folder - Training and testing data set images
- drivinglog.csv - comma seperated values of the dataset
- train_model.ipynb - Model build to train the the car referencing NVIDIA model
- test.py - test the model ona new environment using Anaconda and Unity.
- model.h5 - saved trained model
- environment.yml - Specifying all the requirements
- Final Report
You will need anaconda to run the model.
Creating an environment in Anaconda
conda env create -f environment.yml
running the pre-trained model
python test.py model.h5
To train the model
python train_model.ipynb
To understand the the model architecture in detail go through the Final-Report