Domain-Agnostic Tuning-Encoder for Fast Personalization of Text-To-Image Models

TL;DR

This paper introduces a domain-agnostic encoder-based method for fast personalization of text-to-image models, improving flexibility and semantic fidelity without requiring specialized datasets.

Contribution

The work presents a novel contrastive regularization technique enabling domain-agnostic T2I personalization, surpassing previous methods in flexibility and performance.

Findings

Achieves state-of-the-art personalization performance.

Produces more semantic and flexible token representations.

Does not require specialized datasets or prior concept information.

Abstract

Text-to-image (T2I) personalization allows users to guide the creative image generation process by combining their own visual concepts in natural language prompts. Recently, encoder-based techniques have emerged as a new effective approach for T2I personalization, reducing the need for multiple images and long training times. However, most existing encoders are limited to a single-class domain, which hinders their ability to handle diverse concepts. In this work, we propose a domain-agnostic method that does not require any specialized dataset or prior information about the personalized concepts. We introduce a novel contrastive-based regularization technique to maintain high fidelity to the target concept characteristics while keeping the predicted embeddings close to editable regions of the latent space, by pushing the predicted tokens toward their nearest existing CLIP tokens. Our experimental results demonstrate the effectiveness of our approach and show how the learned tokens are more semantic than tokens predicted by unregularized models. This leads to a better representation that achieves state-of-the-art performance while being more flexible than previous methods.