Weight Scope Alignment: A Frustratingly Easy Method for Model Merging
This addresses model efficiency and robustness issues in applications like federated learning, but it is incremental as it builds on existing averaging-based fusion methods.
The paper tackles the challenge of model merging under training randomness or non-IID data by introducing Weight Scope Alignment (WSA), a simple regularization method that aligns weight scopes to improve merging effectiveness, achieving significant gains in scenarios like Mode Connectivity and Federated Learning.
Merging models becomes a fundamental procedure in some applications that consider model efficiency and robustness. The training randomness or Non-I.I.D. data poses a huge challenge for averaging-based model fusion. Previous research efforts focus on element-wise regularization or neural permutations to enhance model averaging while overlooking weight scope variations among models, which can significantly affect merging effectiveness. In this paper, we reveal variations in weight scope under different training conditions, shedding light on its influence on model merging. Fortunately, the parameters in each layer basically follow the Gaussian distribution, which inspires a novel and simple regularization approach named Weight Scope Alignment (WSA). It contains two key components: 1) leveraging a target weight scope to guide the model training process for ensuring weight scope matching in the subsequent model merging. 2) fusing the weight scope of two or more models into a unified one for multi-stage model fusion. We extend the WSA regularization to two different scenarios, including Mode Connectivity and Federated Learning. Abundant experimental studies validate the effectiveness of our approach.