unigram-tokenization topic

This project shows how to derive the total number of training tokens from a large text dataset from 🤗 datasets with Apache Beam and Dataflow.

sayakpaul