deep learning

To Beam Or Not To Beam: That is a Question of Cooperation for Language GANs

We propose a new training framework for Language GANs based on cooperative decoding search and self-training.