High-quality, multilingual, NLP models. Supports tokenization, lemmatization, part-of-speech tagging, and dependencies