首先，牢記幾個函式，tf.equal()，tf.argmax(array, 1)，tf.reduce_mean()

import tensorflow as tf 
import numpy as np 

A = [1,3,4,5,6]
B = [1,3,4,3,2] 

with tf.Session() as sess: 
    print(sess.run(tf.equal(A, B)))

A == B

[ True  True  True False False]
 False

A 和 B是 list，tensorflow支援list。

import numpy as np

A = [1,3,4,5,6]
B = [1,3,4,3,2] 
A = np.array(A)
B = np.array(B)
A == B
 

array([ True,  True,  True, False, False], dtype=bool)

tf.argmax(array, 1) 指的是 axis=1

tf.argmax(array)預設axis=0

程式碼核心結構如下

images = tf.placeholder(tf.float32, [None, 256, 256, 3], name='input_images')
labels = tf.placeholder(tf.int64, [None], name='labels')

logits, _ = MyNet(images, num_classes=N_CLASSES)
correct_prediction = tf.equal(tf.argmax(logits, 1), labels)
evaluation_step = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))


with tf.Session() as sess:
    init=tf.global_variabels_initializer()
    sess.run(init)   
    validation_accuracy = sess.run(evaluation_step, feed_dict={
                    images: validation_images, labels: validation_labels})
 

numpy 也能計算

X = tf.placeholder(tf.float32,[None,n_dim])
Y = tf.placeholder(tf.float32,[None,n_classes])
    
    
out = one_layer_net(w,b,X)
evaluation_step = np.mean(np.argmax(out,axis=1) == np.argmax(train_Y,axis=1))

with tf.Session() as sess:
    init = tf.global_variables_intializer()
    sess.run(init)
    validation_accuracy = sess.run(evaluation_step, feed_dict={
            X:validation_images,Y:validation_labels})
 

這裡值得一提的是輸入的資料，我們熟悉的輸入資料通常是訓練集的一個batch，而計算accuracy則是把全部資料輸入計算，因此，人們弄出一個驗證集validation set，通常比batch大而比訓練集小。驗證集可以是訓練集中挑選和重複運用，而測試集test set一般不能和訓練集重複。常用的比喻是：驗證集是平時的模擬考試，而測試集是高考。