Linearly Decomposing and Recomposing Vision Transformers for Diverse-Scale Models Shuxia Lin

Open in new window