Professional Cloud Architect – Google Cloud Certification Guide

Training

Now, we'll get to the heart of ML: the actual training. Let's look at the model we have chosen for training. Here, we have X = W * Y + b, where W is weight and b is bias. During training, we will manipulate the W and b values and use the training data to verify whether we are getting correct predictions. Each of these manipulations is referred to as a training step:

While, at the outset, we may obtain poor results, gradually, we will get accurate predictions for the W and b values. If we are not getting the results we anticipated, we may need to change the model:

In the preceding plots, we can see how the line changes when we manipulate ...