TokenVerse: Versatile Multi-concept Personalization in Token Modulation Space

TL;DR

TokenVerse introduces a versatile method for multi-concept personalization in text-to-image diffusion models, enabling disentangled, localized control over complex visual attributes from minimal input images.

Contribution

It offers a novel optimization framework that finds semantic directions in modulation space for multi-concept personalization, handling multiple images and diverse concepts.

Findings

Effective multi-concept disentanglement from few images

Supports complex concept combinations in generated images

Outperforms existing personalization methods

Abstract

We present TokenVerse -- a method for multi-concept personalization, leveraging a pre-trained text-to-image diffusion model. Our framework can disentangle complex visual elements and attributes from as little as a single image, while enabling seamless plug-and-play generation of combinations of concepts extracted from multiple images. As opposed to existing works, TokenVerse can handle multiple images with multiple concepts each, and supports a wide-range of concepts, including objects, accessories, materials, pose, and lighting. Our work exploits a DiT-based text-to-image model, in which the input text affects the generation through both attention and modulation (shift and scale). We observe that the modulation space is semantic and enables localized control over complex concepts. Building on this insight, we devise an optimization-based framework that takes as input an image and a text description, and finds for each word a distinct direction in the modulation space. These directions can then be used to generate new images that combine the learned concepts in a desired configuration. We demonstrate the effectiveness of TokenVerse in challenging personalization settings, and showcase its advantages over existing methods. project's webpage in https://token-verse.github.io/