Running the Model

The final step in the process where the results of all your data collection and training can be evaluated.

Running the model on the Cars

  1. Inside the car’s donkeycar container (~/car directory), run the following command to load the model:

python manage.py drive --model <filepath_to_model>
  1. Open the Web Ui in a web browser by navigating to:

<ip_address_of_edge_device>:8887
  1. Above the camera display, to the left, there is a drop-down that says User, click on it and pick either Full Auto or Auto Steer

    1. Full Auto: gives full control to the AI

    2. Auto Steer: human controls the throttle but the AI controls the steering

Tips for Running the Model

  1. Make sure you are running the model in the same environment you collected data in.

    1. This means more than just running it on the same track. The model will work best if it is also in the same lighting and its surroundings are the same.

Last updated