Model development and training

- [Instructor] After taking care of preparing data sets, we can finally explore model development. It consists of a few steps, model design, training, validation, and testing. In the model design step, we can design one or more model architectures and initialize weights and biases. Usually we take an existing design and modify it. We won't be covering this step in more detail as we are taking care to understand the basics. In model training, we feed the training data into the model, calculate the error, and then adjust the parameters to improve the model's performance. After the training, in the validation step, we measure the model's performance against the data that wasn't used in training. Let's explore how to build our first neural network. We have already imported libraries and data set. Here we define the neural network and we call it net. Then we fill out the init and forward functions. In init, we are adding layers to our network. You can think of them as filters and lenses…