如何在Keras中使用曼哈顿距离

from keras import backend as K
from keras.layers import Layer

class ManhattanDistance(Layer):
    def __init__(self, **kwargs):
        super(ManhattanDistance, self).__init__(**kwargs)

    def call(self, inputs):
        x, y = inputs
        return K.sum(K.abs(x - y), axis=1, keepdims=True)

    def compute_output_shape(self, input_shape):
        return (input_shape[0][0], 1)

input1 = Input(shape=(n,))
input2 = Input(shape=(n,))
distance = ManhattanDistance()([input1, input2])
model = Model(inputs=[input1, input2], outputs=distance)

from keras import backend as K

def manhattan_distance_loss(y_true, y_pred):
    return K.sum(K.abs(y_true - y_pred), axis=1, keepdims=True)

model.compile(loss=manhattan_distance_loss, optimizer='adam')