chunked/nltk_chunking/tasks_text-to-speech/chunk_37.txt

Note that spaces are already replaced by ▁ in the tokenizer and don't need to be handled separately.