texts = ["One document.", "...", "Lots of documents"] for doc in tokenizer.pipe(texts, batch_size=50): pass