ACE-LoRA: Adaptive Orthogonal Decoupling for Continual Image Editing
For practitioners needing diffusion models that adapt to new image editing tasks without forgetting previous ones, this work addresses the underexplored problem of continual learning in image editing.
ACE-LoRA proposes a dynamic regularization framework for continual image editing that mitigates catastrophic forgetting via Adaptive Orthogonal Decoupling and Rank-Invariant Historical Information Compression. It consistently outperforms baselines in instruction fidelity, visual realism, and robustness to forgetting.
State-of-the-art diffusion models often rely on parameter-efficient fine-tuning to perform specialized image editing tasks. However, real-world applications require continual adaptation to new tasks while preserving previously learned knowledge. Despite the practical necessity, continual learning for image editing remains largely underexplored. We propose ACE-LoRA, a dynamic regularization framework for continual image editing that effectively mitigates catastrophic forgetting. ACE-LoRA leverages Adaptive Orthogonal Decoupling to identify and orthogonalize task interference, and introduces a Rank-Invariant Historical Information Compression strategy to address scalability issues in continual updates. To facilitate continual learning in image editing and provide a standardized evaluation protocol, we introduce CIE-Bench, the first comprehensive benchmark in this domain. CIE-Bench encompasses diverse and practically relevant image editing scenarios with a balanced level of difficulty to effectively expose limitations of existing models while remaining compatible with parameter-efficient fine-tuning. Extensive experiments demonstrate that our method consistently outperforms existing baselines in terms of instruction fidelity, visual realism, and robustness to forgetting, establishing a strong foundation for continual learning in image editing.