Implementation of the training framework proposed in Self-Rewarding Language Model, from MetaAI - View it on GitHub