Merging Vision Transformers from Different Tasks and Domains

• Main Authors: Ye, Peng, Huang, Chenyu, Shen, Mingzhu, Chen, Tao, Huang, Yongqi, Zhang, Yuning, Ouyang, Wanli
• Format: Journal Article
• Language: English
• Published: 25.12.2023
• Subjects: Computer Science - Artificial Intelligence; Computer Science - Computer Vision and Pattern Recognition

This work targets to merge various Vision Transformers (ViTs) trained on different tasks (i.e., datasets with different object categories) or domains (i.e., datasets with the same categories but different environments) into one unified model, yielding still good performance on each task or domain. P...