Code for the paper titled Generalization of Finetuned Transformer Language Models to New Clinical Contexts