1 paper
Zhiwei Wang, Jun Huang, Longhua Ma +2
In visual tasks, large teacher models capture essential features and deep information, enhancing performance. However, distilling this information into smaller student models often…