基于SqueezeNet网络的注意力机制.zip

from tensorflow.keras.models import Model from tensorflow.keras.layers import Input,Conv2D,MaxPool2D,GlobalAveragePooling2D,Dropout from tensorflow.keras.layers import concatenate,Dense # from tensorflow.keras.utils import plot_model def fire_model(x, s_1x1, e_1x1, e_3x3, fire_name): # squeeze part squeeze_x = Conv2D(kernel_size=(1,1),filters=s_1x1,padding='same',activation='relu',name=fire_name+'_s1')(x) # expand part expand_x_1 = Conv2D(kernel_size=(1,1),filters=e_1x1,padding='same',activation='relu',name=fire_name+'_e1')(squeeze_x) expand_x_3 = Conv2D(kernel_size=(3,3),filters=e_3x3,padding='same',activation='relu',name=fire_name+'_e3')(squeeze_x) # expand = merge([expand_x_1, expand_x_3], mode='concat', concat_axis=3) expand = concatenate([expand_x_1, expand_x_3],axis=3) return expand def SqueezeNet(nb_classes,inputs=(224, 224,3)): input_img = Input(shape=inputs) conv1 = Conv2D(strides = 2, filters=96, kernel_size=(7,7), padding='same', activation='relu',data_format='channels_last')(input_img) poo1 = MaxPool2D((2,2))(conv1) fire2 = fire_model(poo1, 16, 64, 64,'fire2') fire3 = fire_model(fire2, 16, 64, 64,'fire3') fire4 = fire_model(fire3, 32, 128, 128,'fire4') pool2 = MaxPool2D((2,2))(fire4) fire5 = fire_model(pool2, 32, 128, 128,'fire5') fire6 = fire_model(fire5, 48, 192, 192,'fire6') fire7 = fire_model(fire6, 48, 192, 192,'fire7') fire8 = fire_model(fire7, 64, 256, 256,'fire8') pool3 = MaxPool2D((2,2))(fire8) fire9 = fire_model(pool3, 64, 256, 256,'fire9') dropout1 = Dropout(0.5)(fire9) conv10 = Conv2D(kernel_size=(1,1), filters=100, padding='same', activation='relu')(dropout1) gap = GlobalAveragePooling2D()(conv10) out=Dense(nb_classes,activation='softmax')(gap) return Model(inputs=input_img, outputs=out) if __name__ == "__main__": sq = SqueezeNet(10, inputs=(375, 500, 2)) sq.summary() # plot_model(sq, to_file='model_1d.png', show_shapes=True)