tensorflow作為google開源的項(xiàng)目，現(xiàn)在趕超了caffe，好像成為最受歡迎的深度學(xué)習(xí)框架。確實(shí)在編寫的時(shí)候更能感受到代碼的真實(shí)存在，這點(diǎn)和caffe不同，caffe通過(guò)編寫配置文件進(jìn)行網(wǎng)絡(luò)的生成。環(huán)境tensorflow是0.10的版本，注意其他版本有的語(yǔ)句會(huì)有錯(cuò)誤，這是tensorflow版本之間的兼容問(wèn)題。

還需要安裝PIL：pip install Pillow

圖片的格式： 

– 圖像標(biāo)準(zhǔn)化，可安裝在20×20像素的框內(nèi)，同時(shí)保留其長(zhǎng)寬比。
– 圖片都集中在一個(gè)28×28的圖像中。
– 像素以列為主進(jìn)行排序。像素值0到255，0表示背景（白色），255表示前景（黑色）。

創(chuàng)建一個(gè).png的文件，背景是白色的，手寫的字體是黑色的，

下面是數(shù)據(jù)測(cè)試的代碼，一個(gè)兩層的卷積神經(jīng)網(wǎng)，然后用save進(jìn)行模型的保存。

# coding: UTF-8 import tensorflow as tf import numpy as np import matplotlib.pyplot as plt import input_data ''''' 得到數(shù)據(jù) ''' mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)  training = mnist.train.images trainlable = mnist.train.labels testing = mnist.test.images testlabel = mnist.test.labels  print ("MNIST loaded") # 獲取交互式的方式 sess = tf.InteractiveSession() # 初始化變量 x = tf.placeholder("float", shape=[None, 784]) y_ = tf.placeholder("float", shape=[None, 10]) W = tf.Variable(tf.zeros([784, 10])) b = tf.Variable(tf.zeros([10])) ''''' 生成權(quán)重函數(shù)，其中shape是數(shù)據(jù)的形狀 ''' def weight_variable(shape):   initial = tf.truncated_normal(shape, stddev=0.1)   return tf.Variable(initial) ''''' 生成偏執(zhí)項(xiàng) 其中shape是數(shù)據(jù)形狀 ''' def bias_variable(shape):   initial = tf.constant(0.1, shape=shape)   return tf.Variable(initial)  def conv2d(x, W):   return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')  def max_pool_2x2(x):   return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],              strides=[1, 2, 2, 1], padding='SAME')  W_conv1 = weight_variable([5, 5, 1, 32]) b_conv1 = bias_variable([32]) x_image = tf.reshape(x, [-1, 28, 28, 1])  h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1) h_pool1 = max_pool_2x2(h_conv1)  W_conv2 = weight_variable([5, 5, 32, 64]) b_conv2 = bias_variable([64])  h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2) h_pool2 = max_pool_2x2(h_conv2)   W_fc1 = weight_variable([7 * 7 * 64, 1024]) b_fc1 = bias_variable([1024])  h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64]) h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)  keep_prob = tf.placeholder("float") h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)  W_fc2 = weight_variable([1024, 10]) b_fc2 = bias_variable([10])  y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)  cross_entropy = -tf.reduce_sum(y_*tf.log(y_conv)) train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy) correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1)) accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))  # 保存網(wǎng)絡(luò)訓(xùn)練的參數(shù) saver = tf.train.Saver() sess.run(tf.initialize_all_variables()) for i in range(8000):  batch = mnist.train.next_batch(50)  if i%100 == 0:   train_accuracy = accuracy.eval(feed_dict={     x:batch[0], y_: batch[1], keep_prob: 1.0})   print "step %d, training accuracy %g"%(i, train_accuracy)  train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})  save_path = saver.save(sess, "model_mnist.ckpt") print("Model saved in life:", save_path)  print "test accuracy %g"%accuracy.eval(feed_dict={   x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0})