walkingwindy

` def __get_relative_prob(self, all_close_nei, back_nei_probs): relative_probs = tf.reduce_sum( tf.where( all_close_nei, x=back_nei_probs, y=tf.zeros_like(back_nei_probs), ), axis=1) relative_probs /= tf.reduce_sum(back_nei_probs, axis=1) return relative_probs ` Ahh, I find a tf version of the same...

Thanks for replying. In your implementation, the dimension of numerator is [batchsize] and the dimension of denominator is [batchsize, 1], the broadcasting will result in [batchsize, batchsize] instead of [batchsize,...

Oh！I got it. Thanks for your reply!