TensorFlow技术帖斯坦福TensorFlow教程

tensorflow教程(九)-基本函数使用

2018-06-06  本文已影响79人  致Great

本文主要介绍tf.argmax,tf.reduce_mean(),tf.reduce_sum(),tf.equal()的使用

1 tf.argmax()简介

tf.argmax(vector, 1):返回的是vector中的最大值的索引号,如果vector是一个向量,那就返回一个值,如果是一个矩阵,那就返回一个向量,这个向量的每一个维度都是相对应矩阵行的最大值元素的索引号。

tf.argmax(input=tensor,dimention=axis) 找到给定的张量tensor中在指定轴axis上的最大值/最小值的位置。

实例1

import numpy as np
import tensorflow as tf
A = np.arange(1, 8, 2).reshape(1, 4)
print("A:", A)
B = np.arange(1, 7).reshape(2, 3)
print("B:", B)
with tf.Session() as sess:
    print("A中沿X轴最大值的索引为:", sess.run(tf.argmax(A, 1)))
    print("A中沿Y轴最大值的索引为:", sess.run(tf.argmax(A, 0)))
    print("B中沿X轴最大值的索引为:", sess.run(tf.argmax(B, 1)))
    print("B中沿Y轴最大值的索引为:", sess.run(tf.argmax(B, 0)))

结果


实例2

import tensorflow as tf
a = tf.get_variable(name='a',
                    shape=[3, 4],
                    dtype=tf.float32,
                    initializer=tf.random_uniform_initializer(minval=-1, maxval=1))
b = tf.argmax(input=a, dimension=0)
c = tf.argmax(input=a, dimension=1)
sess = tf.InteractiveSession()
sess.run(tf.initialize_all_variables())
print(sess.run(a))
#[[ 0.04261756 -0.34297419 -0.87816691 -0.15430689]
# [ 0.18663144  0.86972666 -0.06103253  0.38307118]
# [ 0.84588599 -0.45432305 -0.39736366  0.38526249]]
print(sess.run(b))
#[2 1 1 2]
print(sess.run(c))
#[0 1 0]

结果:


实例来自:https://blog.csdn.net/liyaoqing/article/details/54020202

2 tf.reduce_mean()

求平均值:tf.reduce_mean(input_tensor, reduction_indices=None, keep_dims=False, name=None)

实例

import numpy as np
import tensorflow as tf
A = np.arange(0, 6).reshape(2, 3)
print("A:", A)
with tf.Session() as sess:
    print("A中所有值的平均值为:", sess.run(tf.reduce_mean(tf.cast(A, tf.float32))))
    print("A中沿X轴平均值为:", sess.run(tf.reduce_mean(A, 1)))
    print("A中沿Y轴平均值为:", sess.run(tf.reduce_mean(A, 0)))

结果:

A: [[0 1 2]
 [3 4 5]]
A中所有值的平均值为: 2.5
A中沿X轴平均值为: [1 4]
A中沿Y轴平均值为: [1 2 3]
[Finished in 2.3s]

3 tf.reduce_sum()

按某个轴求和

实例

import numpy as np
import tensorflow as tf
A = np.arange(0, 6).reshape(2, 3)
print("A:", A)
with tf.Session() as sess:
    print("A中所有值的和为:", sess.run(tf.reduce_sum(tf.cast(A, tf.float32))))
    print("A中沿X轴和为:", sess.run(tf.reduce_sum(A, 1)))
    print("A中沿Y轴和为:", sess.run(tf.reduce_sum(A, 0)))

结果

A: [[0 1 2]
 [3 4 5]]
A中所有值的和为: 15.0
A中沿X轴和为: [ 3 12]
A中沿Y轴和为: [3 5 7]
[Finished in 2.4s]

4 tf.equal()

tf.equal(real, prediction)是对比这两个矩阵或者向量的相等的元素,如果是相等的那就返回True,反正返回False,返回的值的矩阵维度和real是一样的,我们会在求准确率的时候经常用到它

实例

import tensorflow as tf
import numpy as np

real = [[1, 3, 4, 5, 6]] # 真实值
prediction = [[1, 3, 4, 3, 2]] # 预测值

with tf.Session() as sess:
    correct_preds = tf.equal(real, prediction) #
    print(sess.run(correct_preds))
    correct_preds_num = tf.cast(correct_preds, tf.float32)
    print(sess.run(correct_preds_num))
    accuracy = tf.reduce_mean(correct_preds_num)
    print(sess.run(accuracy))True  True  True False False]]
[[1. 1. 1. 0. 0.]]
0.6
0.6
[Finished in 2.0s]

    # 等同于
    accuracy = tf.reduce_mean(tf.cast(correct_preds, tf.float32))
    print(sess.run(accuracy))

结果:

[[ True  True  True False False]]
[[1. 1. 1. 0. 0.]]
0.6
0.6
[Finished in 2.0s]
上一篇下一篇

猜你喜欢

热点阅读