tensorflow.keras中多输入多输出遇到的几个问题

这里，主要是想构建3个输出，然后计算3个输出的损失函数，并按权重将损失函数加起来作为总的损失

模型构建部分

class MyModel_add3loss(tf.keras.Model):
    def __init__(self):
        super(MyModel_add3loss,self).__init__()

        self.inputa = tf.keras.layers.InputLayer(input_shape=(60, 8, 1))
        self.conv2a = tf.keras.layers.Conv2D(64, (3, 3), padding="same", kernel_initializer="he_normal")#,kernel_regularizer=tf.keras.regularizers.l2(0.0001),bias_regularizer=tf.keras.regularizers.l2(0.0001))
        self.bn_a = tf.keras.layers.BatchNormalization()
        self.relu_a = tf.keras.layers.ReLU()
        self.dropouta = tf.keras.layers.Dropout(0.5)

        self.conv2b = tf.keras.layers.Conv2D(128, (3, 3), padding="same", kernel_initializer="he_normal")#,kernel_regularizer=tf.keras.regularizers.l2(0.0001),bias_regularizer=tf.keras.regularizers.l2(0.0001))
        self.bn_b = tf.keras.layers.BatchNormalization()
        self.relu_b = tf.keras.layers.ReLU()
        self.pool_b = tf.keras.layers.MaxPool2D(strides=(2, 2), padding="same")
        self.dropoutb = tf.keras.layers.Dropout(0.5)

        self.conv2c = tf.keras.layers.Conv2D(256, (3, 3), padding="same", kernel_initializer="he_normal")#,kernel_regularizer=tf.keras.regularizers.l2(0.0001),bias_regularizer=tf.keras.regularizers.l2(0.0001))
        self.bn_c = tf.keras.layers.BatchNormalization()
        self.relu_c = tf.keras.layers.ReLU()
        self.conv2d = tf.keras.layers.Conv2D(256, (3, 3), padding="same", kernel_initializer="he_normal")#,kernel_regularizer=tf.keras.regularizers.l2(0.0001),bias_regularizer=tf.keras.regularizers.l2(0.0001))
        self.bn_d = tf.keras.layers.BatchNormalization()
        self.