Thanks for the thorough tutorial to walk me through. Last bit I don’t fully understand are:
- How do you manage the train/test splitting with the final dataset you came up with yourself?
- Have you done any data augmentation (warping image, luminance varying etc.) in your training?