提出可微自适应合并方法，让模型融合更高效自动。
Merging in a Bottle: Differentiable Adaptive Merging (DAM) and the Path from Averaging to Automation
- 用可微缩放系数动态调整模型权重，实现自适应融合。
- 简单平均法在模型相似时表现不逊于复杂方法。
- 适合需要快速整合多个模型能力的研究者使用。
通过模型合并，AI系统可结合多个语言模型的独特优势，在无需大量重训练的情况下平衡多种能力。然而，由于训练方法和微调差异，整合过程常需专业知识并反复调试。本文探讨从简单平均到自动化方法的融合技术谱系，对比了进化策略、DARE、TIES-Merging等方法。同时提出可微自适应合并（DAM），一种通过缩放系数优化模型集成的高效自适应方法，显著降低计算开销。实验表明，当模型相似度高时，简单平均法如Model Soups仍具竞争力，凸显各类方法的优劣情境。项目已开源，代码与实验流程见GitHub：https://github.com/arcee-ai/DAM。
原文摘要 · Abstract (English)
By merging models, AI systems can combine the distinct strengths of separate language models, achieving a balance between multiple capabilities without requiring substantial retraining. However, the integration process can be intricate due to differences in training methods and fine-tuning, typically necessitating specialized knowledge and repeated refinement. This paper explores model merging techniques across a spectrum of complexity, examining where automated methods like evolutionary strategies stand compared to hyperparameter-driven approaches such as DARE, TIES-Merging and simpler methods like Model Soups. In addition, we introduce Differentiable Adaptive Merging (DAM), an efficient, adaptive merging approach as an alternative to evolutionary merging that optimizes model integration through scaling coefficients, minimizing computational demands. Our findings reveal that even simple averaging methods, like Model Soups, perform competitively when model similarity is high, underscoring each technique's unique strengths and limitations. We open-sourced DAM, including the implementation code and experiment pipeline, on GitHub: https://github.com/arcee-ai/DAM.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。