這是我的自定義填充層：   class CustomZeroPadding2D(Layer):        def __init__(self, **kwargs):            super(CustomZeroPadding2D, self).__init__(**kwargs)        def build(self, input_shape):            super(CustomZeroPadding2D, self).build(input_shape)        def call(self, x):            print('K.int_shape(x)', K.int_shape(x))            print('K.int_shape(K.zeros_like(x))', K.int_shape(K.zeros_like(x)))            res = concatenate([x, K.zeros_like(x)], axis=-1)            return res        def compute_output_shape(self, input_shape):            output_shape = (input_shape[0], input_shape[1], input_shape[2]*2)            return output_shape因為某些原因：K.int_shape(x) (None, 128, 128, 7)但K.int_shape(K.zeros_like(x)) (None, None, None, 7)在doc 中 instantiates an all-zeros variable of the same shape as another tensor，有什么問題？更新：連接不起作用的問題：ValueError: A `Concatenate` layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [(None, 128, 128, 7), (None, None, None, 7)]