DreamCache: Finetuning-Free Lightweight Personalized Image Generation via Feature Caching
This addresses the need for efficient and flexible personalized image generation for users, though it appears incremental as it builds on existing diffusion models with caching and adapters.
The paper tackles the problem of personalized image generation by introducing DreamCache, a method that caches reference image features and uses lightweight adapters to modulate generated images, achieving state-of-the-art alignment with an order of magnitude fewer parameters and improved computational efficiency.
Personalized image generation requires text-to-image generative models that capture the core features of a reference subject to allow for controlled generation across different contexts. Existing methods face challenges due to complex training requirements, high inference costs, limited flexibility, or a combination of these issues. In this paper, we introduce DreamCache, a scalable approach for efficient and high-quality personalized image generation. By caching a small number of reference image features from a subset of layers and a single timestep of the pretrained diffusion denoiser, DreamCache enables dynamic modulation of the generated image features through lightweight, trained conditioning adapters. DreamCache achieves state-of-the-art image and text alignment, utilizing an order of magnitude fewer extra parameters, and is both more computationally effective and versatile than existing models.