Skip to content

4.1 Preparing the model#

Number of masked pixels

How many pixels are masked per patche?

Training dataset dimensions

What are the dimensions of the training set? What is the last dimension?

Configuration

The configuration is printed to the console on the notebook. Yes yes, it is the very long one-liner that starts with N2VConfig.

Here are a few questions:

  1. What do you think "means" and "stds" are and what they are used for?
  2. What is the value of the UNet depth? And what is that?
  3. Can you find the loss name?