PhotorealisticText-to-ImageDiffusionModels withDeepLanguageUnderstanding

While conceptually simple and easy to train, Imagen yields surprisingly strong results. Imagen outperforms other methods on COCO [38] with zero-shot FID-30K of 7.27, significantly outperforming prior work such asGLIDE [43](at 12.4) and the concurrent work ofDALL-E 2[56](at 10.4).