keras自定义层Lambda 输出维度不对

[天下第一小白]

    def reduce_sums(self, vec):
        user_out = vec[0]
        item_out = vec[1]
        res = tf.nn.sigmoid(tf.reduce_sum(tf.multiply(user_out, item_out), axis=1, keep_dims=False))
        return res
    prediction = Lambda(self.reduce_sums)([user_latent, item_latent])

ValueError: Error when checking target: expected lambda_2 to have 1 dimensions, but got array with shape (5167, 1)

lambda_1 (Lambda) (None,) 0

[天下第一小白]

keras自定义层，输出维度总是报错，打印的sumary维度输出没有，没清楚问题出在哪里