Federated learning on csv dataset

I am trying to build a federated learning model but I want to use csv dataset. I have many csv datasets and I want to divide it into clients. How can I make clients and make it distributed ?
Is there any tutorial regarding this ?