Multi-LoRA Composition for Image Generation

TL;DR

This paper introduces two training-free methods for composing multiple LoRAs in image generation, improving the synthesis of complex images by effectively combining diverse LoRAs without additional training.

Contribution

It proposes novel, training-free techniques for multi-LoRA composition and establishes a new benchmark for evaluating their effectiveness in image generation.

Findings

Methods outperform baseline in complex LoRA compositions

Performance improves as more LoRAs are combined

Evaluation with GPT-4V confirms effectiveness

Abstract

Low-Rank Adaptation (LoRA) is extensively utilized in text-to-image models for the accurate rendition of specific elements like distinct characters or unique styles in generated images. Nonetheless, existing methods face challenges in effectively composing multiple LoRAs, especially as the number of LoRAs to be integrated grows, thus hindering the creation of complex imagery. In this paper, we study multi-LoRA composition through a decoding-centric perspective. We present two training-free methods: LoRA Switch, which alternates between different LoRAs at each denoising step, and LoRA Composite, which simultaneously incorporates all LoRAs to guide more cohesive image synthesis. To evaluate the proposed approaches, we establish ComposLoRA, a new comprehensive testbed as part of this research. It features a diverse range of LoRA categories with 480 composition sets. Utilizing an evaluation framework based on GPT-4V, our findings demonstrate a clear improvement in performance with our methods over the prevalent baseline, particularly evident when increasing the number of LoRAs in a composition. The code, benchmarks, LoRA weights, and all evaluation details are available on our project website: https://maszhongming.github.io/Multi-LoRA-Composition.