Saved in:
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.05357 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910727856455680 |
|---|---|
| author | Zhao, Xinyu Sun, Guoheng Cai, Ruisi Zhou, Yukun Li, Pingzhi Wang, Peihao Tan, Bowen He, Yexiao Chen, Li Liang, Yi Chen, Beidi Yuan, Binhang Wang, Hongyi Li, Ang Wang, Zhangyang Chen, Tianlong |
| author_facet | Zhao, Xinyu Sun, Guoheng Cai, Ruisi Zhou, Yukun Li, Pingzhi Wang, Peihao Tan, Bowen He, Yexiao Chen, Li Liang, Yi Chen, Beidi Yuan, Binhang Wang, Hongyi Li, Ang Wang, Zhangyang Chen, Tianlong |
| contents | As Large Language Models (LLMs) excel across tasks and specialized domains, scaling LLMs based on existing models has garnered significant attention, which faces the challenge of decreasing performance when combining disparate models. Various techniques have been proposed for the aggregation of pre-trained LLMs, including model merging, Mixture-of-Experts, and stacking. Despite their merits, a comprehensive comparison and synergistic application of them to a diverse model zoo is yet to be adequately addressed. In light of this research gap, this paper introduces Model-GLUE, a holistic LLM scaling guideline. First, our work starts with a benchmarking of existing LLM scaling techniques, especially selective merging, and variants of mixture. Utilizing the insights from the benchmark results, we formulate an optimal strategy for the selection and aggregation of a heterogeneous model zoo characterizing different architectures and initialization.Our methodology involves the clustering of mergeable models and optimal merging strategy selection, and the integration of clusters through a model mixture. Finally, evidenced by our experiments on a diverse Llama-2-based model zoo, Model-GLUE shows an average performance enhancement of 5.61%, achieved without additional training. Codes are available at: https://github.com/Model-GLUE/Model-GLUE. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_05357 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Model-GLUE: Democratized LLM Scaling for A Large Model Zoo in the Wild Zhao, Xinyu Sun, Guoheng Cai, Ruisi Zhou, Yukun Li, Pingzhi Wang, Peihao Tan, Bowen He, Yexiao Chen, Li Liang, Yi Chen, Beidi Yuan, Binhang Wang, Hongyi Li, Ang Wang, Zhangyang Chen, Tianlong Machine Learning Artificial Intelligence Computation and Language As Large Language Models (LLMs) excel across tasks and specialized domains, scaling LLMs based on existing models has garnered significant attention, which faces the challenge of decreasing performance when combining disparate models. Various techniques have been proposed for the aggregation of pre-trained LLMs, including model merging, Mixture-of-Experts, and stacking. Despite their merits, a comprehensive comparison and synergistic application of them to a diverse model zoo is yet to be adequately addressed. In light of this research gap, this paper introduces Model-GLUE, a holistic LLM scaling guideline. First, our work starts with a benchmarking of existing LLM scaling techniques, especially selective merging, and variants of mixture. Utilizing the insights from the benchmark results, we formulate an optimal strategy for the selection and aggregation of a heterogeneous model zoo characterizing different architectures and initialization.Our methodology involves the clustering of mergeable models and optimal merging strategy selection, and the integration of clusters through a model mixture. Finally, evidenced by our experiments on a diverse Llama-2-based model zoo, Model-GLUE shows an average performance enhancement of 5.61%, achieved without additional training. Codes are available at: https://github.com/Model-GLUE/Model-GLUE. |
| title | Model-GLUE: Democratized LLM Scaling for A Large Model Zoo in the Wild |
| topic | Machine Learning Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2410.05357 |