PixelFlow: Pixel-Space Generative Models with Flow
This addresses the problem of simplifying and improving image generation for AI researchers and practitioners by offering a new paradigm that eliminates the need for pre-trained VAEs.
The paper tackles image generation by introducing PixelFlow, a family of models that operate directly in pixel space instead of using latent-space approaches, achieving an FID of 1.98 on the 256x256 ImageNet class-conditional benchmark.
We present PixelFlow, a family of image generation models that operate directly in the raw pixel space, in contrast to the predominant latent-space models. This approach simplifies the image generation process by eliminating the need for a pre-trained Variational Autoencoder (VAE) and enabling the whole model end-to-end trainable. Through efficient cascade flow modeling, PixelFlow achieves affordable computation cost in pixel space. It achieves an FID of 1.98 on 256$\times$256 ImageNet class-conditional image generation benchmark. The qualitative text-to-image results demonstrate that PixelFlow excels in image quality, artistry, and semantic control. We hope this new paradigm will inspire and open up new opportunities for next-generation visual generation models. Code and models are available at https://github.com/ShoufaChen/PixelFlow.