In [4]:

    

from __future__ import print_function

import tensorflow as tf

# Import MNIST data
from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)

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


    

    




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

In [5]:

    

# tf Graph Input
x = tf.placeholder(tf.float32, [None, 784]) # mnist data image of shape 28*28=784
y = tf.placeholder(tf.float32, [None, 10]) # 0-9 digits recognition => 10 classes

# Set model weights
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))

# Construct model
pred = tf.nn.softmax(tf.matmul(x, W) + b) # Softmax for output probablity

# Minimize error using cross entropy
cost = tf.reduce_mean(-tf.reduce_sum(y*tf.log(pred), reduction_indices=1))
# Gradient Descent
optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)

# Initializing the variables
init = tf.global_variables_initializer()


    

In [10]:

    

# Launch the graph
with tf.Session() as sess:
    sess.run(init)

    # Training cycle
    for epoch in range(training_epochs):
        avg_cost = 0.
        total_batch = int(mnist.train.num_examples/batch_size)
        # Loop over all batches
        for i in range(total_batch):
            batch_xs, batch_ys = mnist.train.next_batch(batch_size)
            # Run optimization op (backprop) and cost op (to get loss value)
            _, c = sess.run([optimizer, cost], feed_dict={x: batch_xs,
                                                          y: batch_ys})
            # Compute average loss
            avg_cost += c / total_batch
        # Display logs per epoch step
        if (epoch+1) % display_step == 0:
            print("Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f}".format(avg_cost))
            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:", accuracy.eval({x: mnist.test.images, y: mnist.test.labels}))


    

    




Epoch: 0001 cost= 1.184215362
Optimization Finished!
Accuracy: 0.8512
Epoch: 0002 cost= 0.665442764
Optimization Finished!
Accuracy: 0.8732
Epoch: 0003 cost= 0.552885519
Optimization Finished!
Accuracy: 0.8821
Epoch: 0004 cost= 0.498674013
Optimization Finished!
Accuracy: 0.8862
Epoch: 0005 cost= 0.465470082
Optimization Finished!
Accuracy: 0.8926
Epoch: 0006 cost= 0.442538116
Optimization Finished!
Accuracy: 0.8941
Epoch: 0007 cost= 0.425505717
Optimization Finished!
Accuracy: 0.8987
Epoch: 0008 cost= 0.412152627
Optimization Finished!
Accuracy: 0.9001
Epoch: 0009 cost= 0.401332922
Optimization Finished!
Accuracy: 0.9018
Epoch: 0010 cost= 0.392403615
Optimization Finished!
Accuracy: 0.9033
Epoch: 0011 cost= 0.384751962
Optimization Finished!
Accuracy: 0.9045
Epoch: 0012 cost= 0.378139112
Optimization Finished!
Accuracy: 0.9065
Epoch: 0013 cost= 0.372421422
Optimization Finished!
Accuracy: 0.9066
Epoch: 0014 cost= 0.367278771
Optimization Finished!
Accuracy: 0.9067
Epoch: 0015 cost= 0.362711139
Optimization Finished!
Accuracy: 0.9084
Epoch: 0016 cost= 0.358637809
Optimization Finished!
Accuracy: 0.9094
Epoch: 0017 cost= 0.354878639
Optimization Finished!
Accuracy: 0.9096
Epoch: 0018 cost= 0.351437983
Optimization Finished!
Accuracy: 0.9106
Epoch: 0019 cost= 0.348301322
Optimization Finished!
Accuracy: 0.9107
Epoch: 0020 cost= 0.345450107
Optimization Finished!
Accuracy: 0.9119
Epoch: 0021 cost= 0.342743502
Optimization Finished!
Accuracy: 0.9121
Epoch: 0022 cost= 0.340244295
Optimization Finished!
Accuracy: 0.9129
Epoch: 0023 cost= 0.337957848
Optimization Finished!
Accuracy: 0.9128
Epoch: 0024 cost= 0.335749747
Optimization Finished!
Accuracy: 0.9142
Epoch: 0025 cost= 0.333680465
Optimization Finished!
Accuracy: 0.9135

In [ ]: