Vector Quantized Diffusion Model for Text-to-Image Synthesis

We present the vector quantized diffusion (VQ-Diffusion) model for text-to-image generation. This method is based on a vector quantized variational autoencoder (VQ-VAE) whose latent space is modeled by a conditional variant of the recently developed Denoising Diffusion Probabilistic Model (DDPM). We find that this latent-space method is well-suited for text-to-image generation tasks because it not only eliminates the unidirectional bias with existing methods but also allows us to incorporate a mask-and-replace diffusion strategy to avoid the accumulation of errors, which is a serious problem with existing methods. Our experiments show that the VQ-Diffusion produces significantly better text-to-image generation results when compared with conventional autoregressive (AR) models with similar numbers of parameters. Compared with previous GAN-based text-to-image methods, our VQ-Diffusion can handle more complex scenes and improve the synthesized image quality by a large margin. Finally, we show that the image generation computation in our method can be made highly efficient by reparameterization. With traditional AR methods, the text-to-image generation time increases linearly with the output image resolution and hence is quite time consuming even for normal size images. The VQ-Diffusion allows us to achieve a better trade-off between quality and speed. Our experiments indicate that the VQ-Diffusion model with the reparameterization is fifteen times faster than traditional AR methods while achieving a better image quality.

PDF Abstract CVPR 2022 PDF CVPR 2022 Abstract

Results from the Paper

 Ranked #1 on Text-to-Image Generation on Oxford 102 Flowers (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Text-to-Image Generation CUB VQ-Diffusion-F FID 10.32 # 5
Text-to-Image Generation CUB VQ-Diffusion-S FID 12.97 # 8
Text-to-Image Generation CUB VQ-Diffusion-B FID 11.94 # 7
Text-to-Image Generation MS COCO VQ-Diffusion-B FID 19.75 # 44
Text-to-Image Generation MS COCO VQ-Diffusion-F FID 13.86 # 38
Text-to-Image Generation Oxford 102 Flowers VQ-Diffusion-F FID 14.1 # 1
Text-to-Image Generation Oxford 102 Flowers VQ-Diffusion-S FID 14.95 # 3
Text-to-Image Generation Oxford 102 Flowers VQ-Diffusion-B FID 14.88 # 2