Dynamic Adapter Meets Prompt Tuning: Parameter-Efficient Transfer Learning for Point Cloud Analysis

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhou, Xin, Liang, Dingkang, Xu, Wei, Zhu, Xingkui, Xu, Yihan, Zou, Zhikang, Bai, Xiang
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916194440708096
author Zhou, Xin
Liang, Dingkang
Xu, Wei
Zhu, Xingkui
Xu, Yihan
Zou, Zhikang
Bai, Xiang
author_facet Zhou, Xin
Liang, Dingkang
Xu, Wei
Zhu, Xingkui
Xu, Yihan
Zou, Zhikang
Bai, Xiang
contents Point cloud analysis has achieved outstanding performance by transferring point cloud pre-trained models. However, existing methods for model adaptation usually update all model parameters, i.e., full fine-tuning paradigm, which is inefficient as it relies on high computational costs (e.g., training GPU memory) and massive storage space. In this paper, we aim to study parameter-efficient transfer learning for point cloud analysis with an ideal trade-off between task performance and parameter efficiency. To achieve this goal, we freeze the parameters of the default pre-trained models and then propose the Dynamic Adapter, which generates a dynamic scale for each token, considering the token significance to the downstream task. We further seamlessly integrate Dynamic Adapter with Prompt Tuning (DAPT) by constructing Internal Prompts, capturing the instance-specific features for interaction. Extensive experiments conducted on five challenging datasets demonstrate that the proposed DAPT achieves superior performance compared to the full fine-tuning counterparts while significantly reducing the trainable parameters and training GPU memory by 95% and 35%, respectively. Code is available at https://github.com/LMD0311/DAPT.
format Preprint
id arxiv_https___arxiv_org_abs_2403_01439
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Dynamic Adapter Meets Prompt Tuning: Parameter-Efficient Transfer Learning for Point Cloud Analysis
Zhou, Xin
Liang, Dingkang
Xu, Wei
Zhu, Xingkui
Xu, Yihan
Zou, Zhikang
Bai, Xiang
Computer Vision and Pattern Recognition
Point cloud analysis has achieved outstanding performance by transferring point cloud pre-trained models. However, existing methods for model adaptation usually update all model parameters, i.e., full fine-tuning paradigm, which is inefficient as it relies on high computational costs (e.g., training GPU memory) and massive storage space. In this paper, we aim to study parameter-efficient transfer learning for point cloud analysis with an ideal trade-off between task performance and parameter efficiency. To achieve this goal, we freeze the parameters of the default pre-trained models and then propose the Dynamic Adapter, which generates a dynamic scale for each token, considering the token significance to the downstream task. We further seamlessly integrate Dynamic Adapter with Prompt Tuning (DAPT) by constructing Internal Prompts, capturing the instance-specific features for interaction. Extensive experiments conducted on five challenging datasets demonstrate that the proposed DAPT achieves superior performance compared to the full fine-tuning counterparts while significantly reducing the trainable parameters and training GPU memory by 95% and 35%, respectively. Code is available at https://github.com/LMD0311/DAPT.
title Dynamic Adapter Meets Prompt Tuning: Parameter-Efficient Transfer Learning for Point Cloud Analysis
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.01439