chunked/nltk_chunking/_trainer/chunk_1.txt

You only need to pass it the necessary pieces for training (model, tokenizer, dataset, evaluation function, training hyperparameters, etc.