This is a text file, not a real PunktSentenceTokenizer model.
Loading it should not work on sentence tokenizer.