Creating a probabilistic generative model to generate synthetic data for a sparse classification problem using TFP

Question

I am wanting to create synthetic data for a sparse classification problem. The idea is that p features used to predict whether or not a sample belongs to one of the C classe