optimization - 优化结果的含义

Question

我运行了 Optuna，发现最好的层数是 5。下面是模型生成代码。我的问题是，5 层是否意味着与其有两层（在循环内），不如有 5 层，脉冲循环外的其余层？或者这是否意味着整个模型中的总层数应该是 5。

def create_model(num_layer, mid_units, num_filters):  

model = Sequential()
model.add( layers.Input( batch_shape=(None,1024,1),dtype=tf.float32,name='audio') )
model.add( layers.Activation('relu') )
model.add(layers.Conv1D(filters = num_filters[0], kernel_size = 2,))
#model.add(layers.MaxPooling1D(pool_size=(2)))
for i in range(1,num_layer):
    model.add(layers.Conv1D(filters = num_filters[i], kernel_size = 2,padding = 'Same',activation ='relu'))
    model.add(layers.MaxPooling1D(pool_size=(2)))
    
   

model.add( layers.MaxPooling1D(pool_size=(4)) )
model.add( layers.Dropout(0.3) )
model.add( layers.Flatten() )
model.add( layers.Dense( mid_units , activation='relu'))
model.add( layers.BatchNormalization() )
model.add( layers.Dropout(0.4))
model.add(layers.Dense( 5 , activation= "softmax" ))
##


return model