--- license: mit --- The BERT model pretrained on the review corpus. Used in the paper ``ReGen: Zero-Shot Text Classification via Training Data Generation with Progressive Dense Retrieval``. See github: https://github.com/yueyu1030/ReGen for details.