Compute corpus BLEU and ROUGE overlap metrics for generated text using the specified token and n-gram rules.
