Fine, I'll Merge It Myself: A Multi-Fidelity Framework for Automated Model Merging
Fuente:
arXiv
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909659511652352 |
|---|---|
| author | Su, Guinan Geiping, Jonas |
| author_facet | Su, Guinan Geiping, Jonas |
| contents | Reasoning capabilities represent a critical frontier for large language models (LLMs), but developing them requires extensive proprietary datasets and computational resources. One way to efficiently supplement capabilities with is by model merging, which offers a promising alternative by combining multiple models without retraining. However, current merging approaches rely on manually-designed strategies for merging hyperparameters, limiting the exploration of potential model combinations and requiring significant human effort. We propose an Automated Model Merging Framework that enables fine-grained exploration of merging strategies while reducing costs through multi-fidelity approximations. We support both single and multi-objective optimization and introduce two novel search spaces: layerwise fusion (LFS) and depth-wise integration (DIS). Evaluating across a number of benchmarks, we find that the search autonomously finds 1) Merges that further boost single-objective performance, even on tasks the model has already been finetuned on, and 2) Merges that optimize multi-objective frontiers across tasks. Effective merges are found with limited compute, e.g. within less than 500 search steps. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_04030 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Fine, I'll Merge It Myself: A Multi-Fidelity Framework for Automated Model Merging Su, Guinan Geiping, Jonas Artificial Intelligence Machine Learning Reasoning capabilities represent a critical frontier for large language models (LLMs), but developing them requires extensive proprietary datasets and computational resources. One way to efficiently supplement capabilities with is by model merging, which offers a promising alternative by combining multiple models without retraining. However, current merging approaches rely on manually-designed strategies for merging hyperparameters, limiting the exploration of potential model combinations and requiring significant human effort. We propose an Automated Model Merging Framework that enables fine-grained exploration of merging strategies while reducing costs through multi-fidelity approximations. We support both single and multi-objective optimization and introduce two novel search spaces: layerwise fusion (LFS) and depth-wise integration (DIS). Evaluating across a number of benchmarks, we find that the search autonomously finds 1) Merges that further boost single-objective performance, even on tasks the model has already been finetuned on, and 2) Merges that optimize multi-objective frontiers across tasks. Effective merges are found with limited compute, e.g. within less than 500 search steps. |
| title | Fine, I'll Merge It Myself: A Multi-Fidelity Framework for Automated Model Merging |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2502.04030 |