efficientnet
efficientnet copied to clipboard
How to take multiple output from pretrained network
input_image = Input(shape=image_input_shape)
eff_model=efn.EfficientNetB3(input_shape=(height, width, 3),
include_top=False,
weights='noisy-student')
model_backbone = Model(eff_model.input,[eff_model.get_layer('block6d_se_squeeze').output,eff_model.output])
timeDistributed_layer = tf.keras.layers.TimeDistributed(model_backbone)(input_image)
t = LSTM(128...)(timeDistributed_layer) -->first output of backbone
c = Conv2D(128,..)(timeDistributed_layer) --> second output of backbone
Based on the sample code above how to specify the output of the backbone to each next network?