What is the train and test data if we want to make word segmentation using machine learning?

Question

I want to use machine learning to apply tokenization. Given a document, I want to generate its tokens. I made a lot of search but I didn\'t find something useful. I am open