Using the technique of "Behavioral Cloning" to train a car to drive itself on a track in Unity3D simulation
Start up the Udacity self-driving simulator, choose lake track and press the Autonomous Mode button. Then, run the model as follows:
python Run_Simulation.py
Fire up the Simulator and record data in training mode for any track.
Run model.ipynb
This will generate files model-<val_loss>.h5
. Choose the file with minimum validation loss as the model file.
Then Run
python Run_Simulation.py --path=[PATH TO MODEL]