chunked/nltk_chunking/_training/chunk_51.txt

This approach works great for smaller datasets, but for larger datasets, you might find it starts to become a problem.