End to End Learning for Self-Driving Golf Carts. All datasets used available here.
Initial code heavily based on Sully Chen's implementation of NVIDIA's End to End Learning for Self-Driving Cars.
To train the end-to-end model, change the dataset path (if need be) in driving_data.py and run:
$ python train.py
To test the model on a local dataset, run:
$ python run_dataset.py
If you want to try and run this model using an external camera in real-time, run:
$ python run.py
If you want to see the ground truth steering angles for a particular dataset visualized, run:
$ python test_dataset.py
Another resource that was extremely helpful, especially in the early days, was udacity-driving-reader for unpacking ROS format datasets. We would also like to thank Udacity for providing training and testing data and for giving advice while working on this project.