Tensorflow MNIST basic

#!/usr/bin/python3

# http://www.tensorfly.cn/tfdoc/tutorials/mnist_beginners.html
# http://www.tensorfly.cn/tfdoc/tutorials/mnist_pros.html

"""
import numpy as np
a = [1,0,0,0,0]
b = [0,1,0,0,0]
c = [0,0,1,0,0]
data = [a,b,c]
data2 = [a,b,a]

print(np.argmax(a)) # 1 dim---> 0 dim
print(np.argmax(b)) # 1 dim---> 0 dim
print(np.argmax(c)) # 1 dim---> 0 dim

#print(np.argmax(data,0)) # 2 dim: matrix---> 1 dim: vector
#print(np.argmax(data2,0))

print(np.argmax(data,1)) # 2 dim: matrix---> 1 dim: vector
print(np.argmax(data2,1))

print( np.equal(np.argmax(data,1),np.argmax(data2,1)) ) # 1 vector


0
1
2
[0 1 2]
[0 1 0]
[ True  True False]
"""

import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
# 默认为0：输出所有log信息
# 设置为1：进一步屏蔽INFO信息
# 设置为2：进一步屏蔽WARNING信息
# 设置为3：进一步屏蔽ERROR信息

import tensorflow as tf
import tensorflow.examples.tutorials.mnist.input_data as input_data

mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)

x = tf.placeholder("float", [None, 784])
y_ = tf.placeholder("float", [None,10]) # real results

W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))

y = tf.nn.softmax(tf.matmul(x,W) + b) # predicted results

cross_entropy = -tf.reduce_sum(y_*tf.log(y)) # cost:  2 matrix ---> scalar

train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)

init = tf.global_variables_initializer()

config_proto = tf.ConfigProto()
config_proto.gpu_options.allow_growth = True # allow gpu dynamic grow
sess = tf.Session(config=config_proto)

sess.run(init) # initial all variables to default 0

with tf.device("/gpu:0"):
    for i in range(1000):
      batch_xs, batch_ys = mnist.train.next_batch(100)
      sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})

    correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1)) # [True,True,False]
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float")) # [1,1,0] ---> 0.67
    print("")
    print(sess.run(accuracy, feed_dict={x: mnist.test.images, y_: mnist.test.labels})) # we got 90.88%