Generating Images with Multimodal Language Models

TL;DR

This paper introduces a multimodal model that combines large language models with visual encoders and decoders, enabling complex image and text generation, retrieval, and dialogue, surpassing previous models in handling longer, more complex language inputs.

Contribution

It presents a novel method to fuse frozen LLMs with visual models via embedding space mapping, allowing flexible multimodal input processing and output generation.

Findings

Outperforms baseline models on complex language tasks

Capable of image retrieval and generation from interleaved inputs

Outperforms non-LLM models in text-to-image tasks

Abstract

We propose a method to fuse frozen text-only large language models (LLMs) with pre-trained image encoder and decoder models, by mapping between their embedding spaces. Our model demonstrates a wide suite of multimodal capabilities: image retrieval, novel image generation, and multimodal dialogue. Ours is the first approach capable of conditioning on arbitrarily interleaved image and text inputs to generate coherent image (and text) outputs. To achieve strong performance on image generation, we propose an efficient mapping network to ground the LLM to an off-the-shelf text-to-image generation model. This mapping network translates hidden representations of text into the embedding space of the visual models, enabling us to leverage the strong text representations of the LLM for visual outputs. Our approach outperforms baseline generation models on tasks with longer and more complex language. In addition to novel image generation, our model is also capable of image retrieval from a prespecified dataset, and decides whether to retrieve or generate at inference time. This is done with a learnt decision module which conditions on the hidden representations of the LLM. Our model exhibits a wider range of capabilities compared to prior multimodal language models. It can process image-and-text inputs, and produce retrieved images, generated images, and generated text -- outperforming non-LLM based generation models across several text-to-image tasks that measure context dependence.