CrossFormer: A Versatile Vision Transformer Hinging on Cross-scale Attention
CrossFormer: A Versatile Vision Transformer Hinging on Cross-scale Attention
整体思路以及计算方式

时间复杂度
训练以及loss
代码
实验以及适用场景
细节
简评
PreviousLocalGlobalNextNested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
Last updated