Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Haixin, Yang, Xinlong, Chang, Jianlong, Jin, Dian, Sun, Jinan, Zhang, Shikun, Luo, Xiao, Tian, Qi
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2305.08381
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916141826310144
author Wang, Haixin
Yang, Xinlong
Chang, Jianlong
Jin, Dian
Sun, Jinan
Zhang, Shikun
Luo, Xiao
Tian, Qi
author_facet Wang, Haixin
Yang, Xinlong
Chang, Jianlong
Jin, Dian
Sun, Jinan
Zhang, Shikun
Luo, Xiao
Tian, Qi
contents Driven by the progress of large-scale pre-training, parameter-efficient transfer learning has gained immense popularity across different subfields of Artificial Intelligence. The core is to adapt the model to downstream tasks with only a small set of parameters. Recently, researchers have leveraged such proven techniques in multimodal tasks and achieve promising results. However, two critical issues remain unresolved: how to further reduce the complexity with lightweight design and how to boost alignment between modalities under extremely low parameters. In this paper, we propose A graceful prompt framework for cross-modal transfer (Aurora) to overcome these challenges. Considering the redundancy in existing architectures, we first utilize the mode approximation to generate 0.1M trainable parameters to implement the multimodal prompt tuning, which explores the low intrinsic dimension with only 0.04% parameters of the pre-trained model. Then, for better modality alignment, we propose the Informative Context Enhancement and Gated Query Transformation module under extremely few parameters scenes. A thorough evaluation on six cross-modal benchmarks shows that it not only outperforms the state-of-the-art but even outperforms the full fine-tuning approach. Our code is available at: https://github.com/WillDreamer/Aurora.
format Preprint
id arxiv_https___arxiv_org_abs_2305_08381
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Parameter-efficient Tuning of Large-scale Multimodal Foundation Model
Wang, Haixin
Yang, Xinlong
Chang, Jianlong
Jin, Dian
Sun, Jinan
Zhang, Shikun
Luo, Xiao
Tian, Qi
Computer Vision and Pattern Recognition
Driven by the progress of large-scale pre-training, parameter-efficient transfer learning has gained immense popularity across different subfields of Artificial Intelligence. The core is to adapt the model to downstream tasks with only a small set of parameters. Recently, researchers have leveraged such proven techniques in multimodal tasks and achieve promising results. However, two critical issues remain unresolved: how to further reduce the complexity with lightweight design and how to boost alignment between modalities under extremely low parameters. In this paper, we propose A graceful prompt framework for cross-modal transfer (Aurora) to overcome these challenges. Considering the redundancy in existing architectures, we first utilize the mode approximation to generate 0.1M trainable parameters to implement the multimodal prompt tuning, which explores the low intrinsic dimension with only 0.04% parameters of the pre-trained model. Then, for better modality alignment, we propose the Informative Context Enhancement and Gated Query Transformation module under extremely few parameters scenes. A thorough evaluation on six cross-modal benchmarks shows that it not only outperforms the state-of-the-art but even outperforms the full fine-tuning approach. Our code is available at: https://github.com/WillDreamer/Aurora.
title Parameter-efficient Tuning of Large-scale Multimodal Foundation Model
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2305.08381