News for hackers. Stuff that matters.

A comment on Lane Following Autopilot with Keras and Tensorflow.

Re:Lane Following Autopilot with Keras and Tensorflow by nilkn on Monday January 23, 2017 @11:57AM (#13463590)
Was the track changed at all during the training? I'm wondering if there's some subtle overfitting here where the car learned to drive along only this specific track. It mentions this but I'm not sure what concrete actions were taken to avoid overfitting:

> The biggest problem I ran into was over fitting the model so that it would not work in evenlly slightly different scenarios.

Regardless, a very cool project.

[ Parent | on HN ]
Re:Lane Following Autopilot with Keras and Tensorflow by wroscoe on Monday January 23, 2017 @02:15PM (#13464995)
The method to avoid overfitting was to use the model with the lowest validation loss, not training loss. I was able to change the track around my house with reasonable success. I think would need many more example turns in the training data to become robust.
[ Parent | on HN ]