[Long Review] Cascaded Diffusion Models for High Fidelity Image Generation скачать в хорошем качестве

[Long Review] Cascaded Diffusion Models for High Fidelity Image Generation

Join 'Speech and Language Technologies' Meetup group https://www.meetup.com/speech-and-lan... to see weekly paper reading schedules and discussions. Cascaded Diffusion Models for High Fidelity Image Generation Jonathan Ho, Chitwan Saharia, William Chan, David J. Fleet, Mohammad Norouzi, Tim Salimans https://arxiv.org/abs/2106.15282 We show that cascaded diffusion models are capable of generating high fidelity images on the class-conditional ImageNet generation benchmark, without any assistance from auxiliary image classifiers to boost sample quality. A cascaded diffusion model comprises a pipeline of multiple diffusion models that generate images of increasing resolution, beginning with a standard diffusion model at the lowest resolution, followed by one or more super-resolution diffusion models that successively upsample the image and add higher resolution details. We find that the sample quality of a cascading pipeline relies crucially on conditioning augmentation, our proposed method of data augmentation of the lower resolution conditioning inputs to the super-resolution models. Our experiments show that conditioning augmentation prevents compounding error during sampling in a cascaded model, helping us to train cascading pipelines achieving FID scores of 1.48 at 64x64, 3.52 at 128x128 and 4.88 at 256x256 resolutions, outperforming BigGAN-deep, and classification accuracy scores of 63.02% (top-1) and 84.06% (top-5) at 256x256, outperforming VQ-VAE-2. #openai #metaai #googleai #diffusion #diffusionmodel #computervision #coursera #ml #course #contrastivelearning #sota #imagegeneration