【TensorFlow】使用TensorFlow实现LR

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

load data

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("/tmp/data/", one_hot=True)

Extracting /tmp/data/train-images-idx3-ubyte.gz
Extracting /tmp/data/train-labels-idx1-ubyte.gz
Extracting /tmp/data/t10k-images-idx3-ubyte.gz
Extracting /tmp/data/t10k-labels-idx1-ubyte.gz

参数

learning_rate = 0.01
training_epochs = 25
batch_size = 100
display_step = 1

定义模型

# 计算图的输入 
X = tf.placeholder(tf.float32,[None,784]) # mnist图片尺寸为28*28=784
Y = tf.placeholder(tf.float32,[None,10]) # 0-9共9个数字，10分类问题
# 模型权重
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))
# 构建模型
pred = tf.nn.softmax(tf.matmul(X,W)+b)
# crossentroy
loss = tf.reduce_mean(-tf.reduce_sum(Y*tf.log(pred), reduction_indices=1))
# SGD
optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(loss)
# 初始化
init = tf.global_variables_initializer()

训练

with tf.Session() as sess:
    sess.run(init)
    
    for epoch in range(training_epochs):
        avg_loss = 0.
        total_batch = int(mnist.train.num_examples/batch_size)
        
        for i in range(total_batch):
            batch_xs, batch_ys = mnist.train.next_batch(batch_size)
            _,l = sess.run([optimizer,loss],feed_dict={X:batch_xs,Y:batch_ys})
            avg_loss += l / total_batch
            
        if (epoch+1)%display_step == 0:
            print("Epoch:", '%04d' % (epoch+1), "loss=", "{:.9f}".format(avg_loss))
    
    print("Optimization Finished!")
    
     # Test model
    correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(Y, 1))
    # Calculate accuracy
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    #print("Accuracy:",sess.run([accuracy],feed_dict={X: mnist.test.images, Y: mnist.test.labels}))
    print("Accuracy:", accuracy.eval({X: mnist.test.images, Y: mnist.test.labels}))

Epoch: 0001 loss= 1.183849382
Epoch: 0002 loss= 0.665282110
Epoch: 0003 loss= 0.552660186
Epoch: 0004 loss= 0.498592658
Epoch: 0005 loss= 0.465719688
Epoch: 0006 loss= 0.442414914
Epoch: 0007 loss= 0.425411000
Epoch: 0008 loss= 0.412160496
Epoch: 0009 loss= 0.401395797
Epoch: 0010 loss= 0.392373819
Epoch: 0011 loss= 0.385098133
Epoch: 0012 loss= 0.377928982
Epoch: 0013 loss= 0.372355448
Epoch: 0014 loss= 0.367128593
Epoch: 0015 loss= 0.362637794
Epoch: 0016 loss= 0.358807674
Epoch: 0017 loss= 0.354965761
Epoch: 0018 loss= 0.351236291
Epoch: 0019 loss= 0.348444075
Epoch: 0020 loss= 0.345302655
Epoch: 0021 loss= 0.343172931
Epoch: 0022 loss= 0.339864279
Epoch: 0023 loss= 0.338058026
Epoch: 0024 loss= 0.335610659
Epoch: 0025 loss= 0.333698551
Optimization Finished!
Accuracy: [0.9137]