机器学习入门--手写体识别

一.python3.5下载（跳过)

下载地址：
[https://www.python.org/downloads/release/python-350/]
安装方法：
[https://jingyan.baidu.com/article/29697b9158e688ab21de3c75.html]

二.下载anaconda（自带python 3.5）

最好下载3-4.2.0
下载地址
安装方法

三.下载安装tensorflow

在anaconda prompt 里敲命令安装tensorflow(一定要用管理员打开anaconda prompt不然可能报cmd 不是内部指令 若无法解决自行百度)
安装方法
https://www.cnblogs.com/ming-4/p/11516728.html
如果出现错误
解决办法
如果该解决方法里的命令不对，看他的评论
然后又遇到了

查看python版本和pip版本

实在不行，考虑换镜像。

pip install https://mirrors.tuna.tsinghua.edu.cn/tensorflow/windows/cpu/tensorflow-1.1.0-cp35-cp35m-win_amd64.whl

作者链接

很明显我是小白，走到弯路不比你们少。
（到这的步骤已经搞了一上午）
用Spyder 测试Tensorflow安装成功(感谢凡哥提供,转载于 悲恋花丶无心之人)
如果进入Spyder import tensorflow 没找到 tensorflow 点击参照方法二

四.python 手写体识别

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

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'

# 定义一个全局对象来获取参数的值，在程序中使用(eg：FLAGS.iteration)来引用参数
FLAGS = tf.app.flags.FLAGS

# 设置训练相关参数
tf.app.flags.DEFINE_integer("iteration", 10001, "Iterations to train [1e4]")
tf.app.flags.DEFINE_integer("disp_freq", 200, "Display the current results every display_freq iterations [1e2]")
tf.app.flags.DEFINE_integer("train_batch_size", 100, "The size of batch images [128]")
tf.app.flags.DEFINE_float("learning_rate", 0.1, "Learning rate of for adam [0.01]")
tf.app.flags.DEFINE_string("log_dir", "logs", "Directory of logs.")


def main(argv=None):
    # 0、准备训练/验证/测试数据集
    mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

    # 1、数据输入设计：使用 placeholder 将数据送入网络，None 表示张量的第一个维度可以是任意长度的
    with tf.name_scope('Input'):
        X = tf.placeholder(dtype=tf.float32, shape=[None, 784], name='X_placeholder')
        Y = tf.placeholder(dtype=tf.int32, shape=[None, 10], name='Y_placeholder')

    # 2、前向网络设计
    with tf.name_scope('Inference'):
        W = tf.Variable(initial_value=tf.random_normal(shape=[784, 10], stddev=0.01), name='Weights')
        b = tf.Variable(initial_value=tf.zeros(shape=[10]), name='bias')
        logits = tf.matmul(X, W) + b
        Y_pred = tf.nn.softmax(logits=logits)

    # 3、损失函数设计
    with tf.name_scope('Loss'):
        # 求交叉熵损失
        cross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=Y, logits=logits, name='cross_entropy')
        # 求平均
        loss = tf.reduce_mean(cross_entropy, name='loss')

    # 4、参数学习算法设计
    with tf.name_scope('Optimization'):
        optimizer = tf.train.GradientDescentOptimizer(FLAGS.learning_rate).minimize(loss)

    # 5、评估节点设计
    with tf.name_scope('Evaluate'):
        # 返回验证集/测试集预测正确或错误的布尔值
        correct_prediction = tf.equal(tf.argmax(Y_pred, 1), tf.argmax(Y, 1))
        # 将布尔值转换为浮点数后，求平均准确率
        accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

    print('~~~~~~~~~~~开始执行计算图~~~~~~~~~~~~~~')
    with tf.Session() as sess:
        summary_writer = tf.summary.FileWriter(logdir=FLAGS.log_dir, graph=sess.graph)
        # 初始化所有变量
        sess.run(tf.global_variables_initializer())
        total_loss = 0
        for i in range(0, FLAGS.iteration):
            X_batch, Y_batch = mnist.train.next_batch(FLAGS.train_batch_size)
            _, loss_batch = sess.run([optimizer, loss], feed_dict={X: X_batch, Y: Y_batch})
            total_loss += loss_batch
            if i % FLAGS.disp_freq == 0:
                val_acc = sess.run(accuracy, feed_dict={X: mnist.validation.images, Y: mnist.validation.labels})
                if i == 0:
                    print('step: {}, train_loss: {}, val_acc: {}'.format(i, total_loss, val_acc))
                else:
                    print('step: {}, train_loss: {}, val_acc: {}'.format(i, total_loss / FLAGS.disp_freq, val_acc))
                total_loss = 0

        test_acc = sess.run(accuracy, feed_dict={X: mnist.test.images, Y: mnist.test.labels})
        print('test accuracy: {}'.format(test_acc))
        summary_writer.close()


# 执行main函数
if __name__ == '__main__':
    tf.app.run()

# -*- coding: utf-8 -*-
"""
Created on Mon Feb 24 16:18:26 2020

TREEGER

@author: Administrator
"""

import tensorflow as tf
import ssl

ssl._create_default_https_context = ssl._create_unverified_context
from tensorflow.examples.tutorials.mnist import input_data

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

learning_rate = 0.005
training_epochs = 20
batch_size = 100
batch_count = int(mnist.train.num_examples / batch_size)
n_hidden_1 = 256
n_hidden_2 = 256
n_input = 784
n_classes = 10  # (0-9 数字)
X = tf.placeholder("float", [None, n_input])
Y = tf.placeholder("float", [None, n_classes])

weights = {
    'weight1': tf.Variable(tf.random_normal([n_input, n_hidden_1])),
    'weight2': tf.Variable(tf.random_normal([n_hidden_1, n_hidden_2])),
    'out': tf.Variable(tf.random_normal([n_hidden_2, n_classes]))
}
biases = {
    'bias1': tf.Variable(tf.random_normal([n_hidden_1])),
    'bias2': tf.Variable(tf.random_normal([n_hidden_2])),
    'out': tf.Variable(tf.random_normal([n_classes]))
}


def multilayer_perceptron_model(x):
    layer_1 = tf.add(tf.matmul(x, weights['weight1']), biases['bias1'])
    layer_2 = tf.add(tf.matmul(layer_1, weights['weight2']), biases['bias2'])
    out_layer = tf.matmul(layer_2, weights['out']) + biases['out']
    return out_layer


logits = multilayer_perceptron_model(X)
loss_op = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=Y))
optimizer = tf.train.GradientDescentOptimizer(learning_rate)
# optimizer = tf.train.MomentumOptimizer(learning_rate,0.2)
# optimizer = tf.train.AdagradOptimizer(learning_rate)
# optimizer = tf.train.AdamOptimizer(learning_rate)
train_op = optimizer.minimize(loss_op)
init = tf.global_variables_initializer()  # 参数初始化

with tf.Session() as sess:
    sess.run(init)
    for epoch in range(training_epochs):  # range(150):training_epochs
        avg_cost = 0.
        for i in range(batch_count):
            train_x, train_y = mnist.train.next_batch(batch_size)
            _, c = sess.run([train_op, loss_op], feed_dict={X: train_x, Y: train_y})
            avg_cost += c / batch_count
        print("Epoch:", '%02d' % (epoch + 1), "avg cost={:.6f}".format(avg_cost))

        pred = tf.nn.softmax(logits)  # Apply softmax to logits
        correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(Y, 1))
        accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
        print("Accuracy:", accuracy.eval({X: mnist.test.images, Y: mnist.test.labels}))