Advancing Vision-Language Models with Adapter Ensemble Strategies
Yue Bai, Handong Zhao, Zhe Lin, Ajinkya Kale, Jiuxiang Gu, Tong Yu, Sungchul Kim, Yun Fu
Abstract
CLIP revolutes vision-language pretraining by using contrastive learning on paired web data. However, the sheer size of these pretrained models makes full-model finetuning exceedingly costly. One common solution is the “adapter”, which finetunes a few additional parameters while freezing the backbone. It harnesses the heavy-duty backbone while offering a light finetuning for small downstream tasks. This synergy prompts us to explore the potential of augmenting large-scale backbones with traditional machine learning techniques. Often employed in traditional fields and overlooked in the large-scale era, these techniques could provide valuable enhancements. Herein, we delve into the “adapter ensembles” in the realm of large-scale pretrained vision-language models. We begin with a proof-of-concept study to establish the efficacy of combining multiple adapters. We then present extensive evidence showing these ensembles excel in a variety of settings, particularly when employing a Multi-Scale Attention (MSA) approach thoughtfully integrated into the ensemble framework. We further incorporate the LoRA to mitigate the additional parameter burden. We focus on vision-language retrieval, using different backbones under constraints of minimal data, parameters, and finetuning budgets. This research paves the way for a synergistic blend of traditional, yet effective, strategies with modern large-scale networks.- Anthology ID:
- 2024.findings-emnlp.921
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2024
- Month:
- November
- Year:
- 2024
- Address:
- Miami, Florida, USA
- Editors:
- Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 15702–15720
- Language:
- URL:
- https://aclanthology.org/2024.findings-emnlp.921
- DOI:
- 10.18653/v1/2024.findings-emnlp.921
- Cite (ACL):
- Yue Bai, Handong Zhao, Zhe Lin, Ajinkya Kale, Jiuxiang Gu, Tong Yu, Sungchul Kim, and Yun Fu. 2024. Advancing Vision-Language Models with Adapter Ensemble Strategies. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 15702–15720, Miami, Florida, USA. Association for Computational Linguistics.
- Cite (Informal):
- Advancing Vision-Language Models with Adapter Ensemble Strategies (Bai et al., Findings 2024)
- PDF:
- https://preview.aclanthology.org/dois-2013-emnlp/2024.findings-emnlp.921.pdf