IterIS: Iterative Inference-Solving Alignment for LoRA Merging

TL;DR

IterIS introduces an iterative inference-based optimization method for merging LoRA adapters, significantly reducing data requirements and improving performance across various large model tasks while ensuring efficiency.

Contribution

The paper presents a novel optimization framework with iterative refinement, regularization, and adaptive weighting for LoRA merging, addressing limitations of prior methods.

Findings

Outperforms existing LoRA merging methods on multiple benchmarks.

Requires only 1-5% of unlabeled samples compared to prior approaches.

Achieves convergence with minimal steps, ensuring efficiency.

Abstract

Low-rank adaptations (LoRA) are widely used to fine-tune large models across various domains for specific downstream tasks. While task-specific LoRAs are often available, concerns about data privacy and intellectual property can restrict access to training data, limiting the acquisition of a multi-task model through gradient-based training. In response, LoRA merging presents an effective solution by combining multiple LoRAs into a unified adapter while maintaining data privacy. Prior works on LoRA merging primarily frame it as an optimization problem, yet these approaches face several limitations, including the rough assumption about input features utilized in optimization, massive sample requirements, and the unbalanced optimization objective. These limitations can significantly degrade performance. To address these, we propose a novel optimization-based method, named IterIS: 1) We formulate LoRA merging as an advanced optimization problem to mitigate the rough assumption. Additionally, we employ an iterative inference-solving framework in our algorithm. It can progressively refine the optimization objective for improved performance. 2) We introduce an efficient regularization term to reduce the need for massive sample requirements (requiring only 1-5% of the unlabeled samples compared to prior methods). 3) We utilize adaptive weights in the optimization objective to mitigate potential unbalances in LoRA merging process. Our method demonstrates significant improvements over multiple baselines and state-of-the-art methods in composing tasks for text-to-image diffusion, vision-language models, and large language models. Furthermore, our layer-wise algorithm can achieve convergence with minimal steps, ensuring efficiency in both memory and computation.