For language generation, AFAIK there is no good model that follows this architecture. For image generation, Generative Adversarial Networks are strong contender

For language generation, AFAIK there is no good model that follows this architecture. For image generation, Generative Adversarial Networks are strong contenders. See for instance: https://github.com/Newmu/dcgan_code https://github.com/Newmu/dcgan_code