NLG, Explainable Transformers, Text Evaluation