Creating a vector quantizer

You can use neural networks for vector quantization as well. Vector quantization is the N-dimensional version of rounding off. This is very commonly used across multiple areas in computer vision, NLP, and machine learning in general.

Getting ready

In previous recipes, we have already addressed vector quantization concepts: Compressing an image using vector quantization and Creating features using visual Codebook and vector quantization. In this recipe, we will define a neural network with two layers—10 neurons in input layer and 4 neurons in the output layer. Then we will use this network to divide the space into four regions.

Before starting, you need to make a change to fix a library bug...