Shulou(Shulou.com)06/02 Report--

Today, I will talk to you about Softmax logical regression in TensorFlow about how to identify handwritten digital MNIST data sets, which may not be well understood by many people. in order to make you understand better, the editor has summarized the following contents for you. I hope you can get something according to this article.

Logical regression model based on MNIST dataset to do ten-class task

Softmax Regression without hidden layer can only infer which number it is directly from the pixels of the image, and there is no process of feature abstraction. Depending on the hidden layer, the multi-layer neural network can combine high-order features, such as horizontal lines, vertical lines, circles and so on, and then these high-order features or components can be combined into numbers to achieve accurate matching and classification.

Import tensorflow as tfimport numpy as npimport input_dataprint ('Download and Extract MNIST dataset') mnist = input_data.read_data_sets (' data/') One_hot=True) # one_hot=True means encoding print ("tpye of 'mnist' is% s"% (type (mnist) print ("number of train data is% d"% (mnist.train.num_examples)) print ("number of test data is% d"% (mnist.test.num_examples)) trainimg = mnist.train.imagestrainlabel = mnist.train.labelstestimg = mnist.test.imagestestlabel = mnist.test.labelsprint ("MNIST loaded") ) "" print ("type of 'trainimg' is% s"% (type (trainimg) print ("type of' trainlabel' is% s"% (type (trainlabel) print ("type of 'testimg' is% s"% (type (testimg) print ("type of' testlabel' is% s"% (type (testlabel) print ("- -- ") print (" shape of 'trainimg' is% s "% (trainimg.shape) ) print ("shape of 'trainlabel' is%"% (trainlabel.shape,)) print ("shape of' testimg' is%"% (testimg.shape,)) print ("shape of 'testlabel' is% s"% (testlabel.shape,)) "" x = tf.placeholder (tf.float32, [None, 784]) y = tf.placeholder (tf.float32, [None, 10]) # None is for infinitew = tf.Variable (tf.zeros ([784]) 10])) # for convenience, initialize directly with 0 Tasks classified by b = tf.Variable (tf.zeros ([10])) # 10 can be initialized by Gauss, 10 kinds of label So you only need to initialize 10 bpred = tf.nn.softmax (tf.matmul (x, w) + b) # predicted values of forward propagation cost = tf.reduce_mean (- tf.reduce_sum (y*tf.log (pred), reduction_indices= [1])) # Cross Entropy loss function optm = tf.train.GradientDescentOptimizer (0.01) .minimize (cost) corr = tf.equal (tf.argmax (pred, 1), tf.argmax (y) 1) # tf.equal () compare whether the predicted index is the same as the real label index Return True as well Different returns Falseaccr = tf.reduce_mean (tf.cast (corr) Tf.float32) init = tf.global_variables_initializer () # Global parameter initializer training_epochs = 100all sample iterations batch_size = 100# Select 100samples per iteration display_step = examples SESSIONsess = tf.Session () # define a Sessionsess.run (init) # run initialization operation in sess # MINI-BATCH LEARNINGfor epoch in range (training_epochs): # per An epoch loops avg_cost = 0. # at the beginning, the loss value is defined as 0 num_batch = int (mnist.train.num_examples/batch_size) for i in range (num_batch): # each batch selects batch_xs, batch_ys = mnist.train.next_batch (batch_size) # you can get data by batch through next_batch () Sess.run (optm, feed_dict= {x: batch_xs, y:batch_ys}) # run is solved by gradient descent, and XMagi y is passed in through placeholder avg_cost + = sess.run (cost, feed_dict= {x: batch_xs, y:batch_ys}) / num_batch # DISPLAY if epoch% display_step = = 0: # display_step was previously defined as 5 Print every 5 epoch here train_acc = sess.run (accr, feed_dict= {x: batch_xs, y:batch_ys}) test_acc = sess.run (accr, feed_dict= {x: mnist.test.images, y: mnist.test.labels}) print ("Epoch: d cost:% .9f TRAIN ACCURACY:% .3f TEST ACCURACY:% .3f"% (epoch, training_epochs, avg_cost) Train_acc, test_acc)) print ("DONE")

Iterate 100 times to run the model, and finally, in the test set can achieve 92.2% accuracy, although not bad, but not to the degree of practical. The main application scenario of handwritten number recognition is to identify bank checks, which may cause serious consequences if the accuracy is not high enough.

Epoch: 095/100 loss: 0.283259882 train_acc: 0.940 test_acc: 0.922

Insert some knowledge points about the use of some functions in tensorflow

Sess = tf.InteractiveSession () arr = np.array ([31, 23, 4, 24, 27, 34], [18, 3, 25, 0, 6, 35], [28, 14, 33, 22, 30, 8], [13, 30, 21, 19, 7, 9], [16, 1, 26, 32, 2, 29] [17, 12, 5, 11, 10, 15]) print in tensorflow using .eval () tf.rank (arr). Eval () # print the dimension of matrix arr tf.shape (arr). Eval () # print the size of matrix arr tf.argmax (arr, 0). Eval () # print the index of the maximum value Parameter 0 is indexed by column and 1 is indexed by row. Do you have any further understanding of how Softmax logical regression in TensorFlow identifies handwritten numeral MNIST datasets? If you want to know more knowledge or related content, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.