张量流:在矩阵乘法和加法期间为纳米

时间:2017-06-06 12:19:12

标签: python matrix tensorflow nan

我使用函数(ndegree_poly)获取张量和权重数组,并从中计算多项式的结果。

代码似乎很直接,但是当度数增加,或者函数重复多次时,结果张量包含一堆nans和infs。 infs是合理的。但如果数字变得非常小,那么它们不应该变为零,而不是纳米?

import tensorflow as tf

function_degree = 10

def ndegree_poly(x, a, degree=6):
        op = tf.add_n([tf.multiply(tf.pow(x, i), a[i]) for i in range(1, degree)])
        return tf.add(op, a[0])

with tf.Session() as sess:
    poly_weight = tf.Variable(tf.random_normal([function_freedom, 1, 5]))

    mat = tf.Variable(tf.random_normal([2, 5]))

    result0 = ndegree_poly(mat, poly_weight, function_degree)
    result1 = ndegree_poly(result0, poly_weight, function_degree)
    result2 = ndegree_poly(result1, poly_weight, function_degree)
    result3 = ndegree_poly(result2, poly_weight, function_degree)
    result4 = ndegree_poly(result3, poly_weight, function_degree)


    sess.run(tf.global_variables_initializer())
    print(sess.run(result4))

打印:

[[-0.28569764         nan         nan         nan        -inf]
 [        nan         nan  3.55561209         nan  0.53827095]]

1 个答案:

答案 0 :(得分:1)

nan值不是来自非常小的系数,它只是尝试做∞ - ∞的“自然”结果,因为系数来自正态分布,因此都是正数和负数。 / p>

import math
import tensorflow as tf

tf_inf = tf.constant(inf)
res = tf_inf - tf_inf
with tf.Session() as sess:
    print(sess.run(res))

>>> nan