JetFormer: An Autoregressive Generative Model of Raw Images and Text
This work addresses the need for streamlined multimodal models in AI by removing reliance on separately trained components, though it appears incremental as it builds on existing transformer and normalizing flow techniques.
The authors tackled the problem of joint generative modeling of images and text by proposing JetFormer, an autoregressive decoder-only transformer that directly maximizes raw data likelihood without separate pretrained components, achieving text-to-image generation quality competitive with recent VQ-VAE- and VAE-based baselines and demonstrating robust image understanding capabilities.
Removing modeling constraints and unifying architectures across domains has been a key driver of the recent progress in training large multimodal models. However, most of these models still rely on many separately trained components such as modality-specific encoders and decoders. In this work, we further streamline joint generative modeling of images and text. We propose an autoregressive decoder-only transformer - JetFormer - which is trained to directly maximize the likelihood of raw data, without relying on any separately pretrained components, and can understand and generate both text and images. Specifically, we leverage a normalizing flow model to obtain a soft-token image representation that is jointly trained with an autoregressive multimodal transformer. The normalizing flow model serves as both an image encoder for perception tasks and an image decoder for image generation tasks during inference. JetFormer achieves text-to-image generation quality competitive with recent VQ-VAE- and VAE-based baselines. These baselines rely on pretrained image autoencoders, which are trained with a complex mixture of losses, including perceptual ones. At the same time, JetFormer demonstrates robust image understanding capabilities. To the best of our knowledge, JetFormer is the first model that is capable of generating high-fidelity images and producing strong log-likelihood bounds.