Rethinking Super-Resolution as Text-Guided Details Generation
Please refer to the paper AttnGAN: Fine-Grained Text to Image Generation with Attentional Generative Adversarial Networks for training the DAMSMencoders.