#1:導入所需的軟件 import tensorflow as tf ''' 獲取mnist數據放在當前文件夾下，利用input_data函數解析該數據集 train_img和train——label構成訓練集，包含60000個手寫體數字圖片和對應的標簽 test_img和test_label表示測試集，包含10000個樣本和10000個標簽 ''' from tensorflow.examples.tutorials.mnist import input_data mnist = input_data.read_data_sets('MNIST_data/',one_hot=True)train_img = mnist.train.images train_label = mnist.train.labels test_img = mnist.test.images test_label = mnist.test.labels#3:在Tensorflow圖中為訓練數據集的輸入x和標簽y創建占位符x = tf.compat.v1.placeholder(tf.float32,[None,784],name='X') y = tf.compat.v1.placeholder(tf.float32,[None,10],name='Y')#4:創建學習變量，權重和偏置W = tf.compat.v1.Variable(tf.zeros([784,10]),name='W') b = tf.compat.v1.Variable(tf.zeros([10]),name='b')#5:創建邏輯回歸模型。y_hat = tf.nn.softmax(tf.matmul(x,W)+b)#6: 損失（loss）函數loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=y_hat))#7:采用Tensorflow GradientDescentOptimizer,學習率為0.01optimizer = tf.compat.v1.train.GradientDescentOptimizer(learning_rate=0.01).minimize(loss)#預測結果 pred = tf.equal(tf.argmax(y_hat,1),tf.argmax(y,1))#計算準確率 accuracy = tf.reduce_mean(tf.cast(pred,'float'))#8:為變量進行初始化： init = tf.global_variables_initializer() training_epochs = 50 batch_size = 100 display_step = 5with tf.compat.v1.Session() as sess:sess.run(init)summary_writer = tf.summary.FileWriter('graphs3',sess.graph)for epoch in range(training_epochs):loss_avg = 0num_of_batch = int(mnist.train.num_examples/batch_size)for i in range(num_of_batch):batch_xs,batch_ys = mnist.train.next_batch(batch_size)feeds_train = {x:batch_xs,y:batch_ys}sess.run([optimizer,loss],feed_dict=feeds_train)loss_avg+=sess.run(loss,feed_dict=feeds_train)/num_of_batch#訓練過程輸出if epoch % display_step ==0:feeds_test = {x:mnist.test.images,y:mnist.test.labels}train_acc = sess.run(accuracy,feed_dict=feeds_train)test_acc = sess.run(accuracy,feed_dict=feeds_test)print()print('Epoch {0}:Loss {1}:train_acc:{2}:test_acc:{3}'.format(epoch,loss_avg,train_acc,test_acc))print("Done")

結果輸入

總結

以上是生活随笔為你收集整理的Tensorflow逻辑回归处理MNIST数据集的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。