Open XcloudFance opened 2 months ago
I havent seen any part of the paper mentioned about it. But the source code contains SENet in patchembed layers.
Regarding of my concerns, are those results shown in the paper also benefited from Channel Attention? Are the model parameters also involving SENet?
Sorry, it is in patchmerging layer, not patchembed layer
I havent seen any part of the paper mentioned about it. But the source code contains SENet in patchembed layers.
Regarding of my concerns, are those results shown in the paper also benefited from Channel Attention? Are the model parameters also involving SENet?