size_categories: - 1M<n<10M
A dataset of 8 million sentences in hundreds of languages, crowd sourced from tatoeba.