Knowledge Fusion By Evolving Weights of Language Models

TL;DR

This paper introduces Evolver, a novel evolutionary algorithm-based method for fusing multiple pre-trained language models into a unified, high-performing model without additional training, demonstrating superior results across various benchmarks.

Contribution

Evolver is a new knowledge fusion technique that combines models via weight evolution, improving performance without extra training or data, and integrates easily with existing frameworks.

Findings

Evolver outperforms state-of-the-art models on mainstream benchmarks.

The method effectively generalizes across different data domains.

It can be applied to various types of language models.

Abstract

Fine-tuning pre-trained language models, particularly large language models, demands extensive computing resources and can result in varying performance outcomes across different domains and datasets. This paper examines the approach of integrating multiple models from diverse training scenarios into a unified model. This unified model excels across various data domains and exhibits the ability to generalize well on out-of-domain data. We propose a knowledge fusion method named Evolver, inspired by evolutionary algorithms, which does not need further training or additional training data. Specifically, our method involves aggregating the weights of different language models into a population and subsequently generating offspring models through mutation and crossover operations. These offspring models are then evaluated against their parents, allowing for the preservation of those models that show enhanced performance on development datasets. Importantly, our model evolving strategy can be seamlessly integrated with existing model merging frameworks, offering a versatile tool for model enhancement. Experimental results on mainstream language models (i.e., encoder-only, decoder-only, encoder-decoder) reveal that Evolver outperforms previous state-of-the-art models by large margins. The code is publicly available at {https://github.com/duguodong7/model-evolution}.