머신러닝스터디/2016/2016_05_14

# python
import tensorflow as tf
import input_data

if __name__ == "__main__":

    # Logistic Regression
    # use sigmoid

    # Input
    # MNIST
    # Cross Entropy
    mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)
    learning_rate = 0.01
    x = tf.placeholder(tf.float32, shape=(None, 28 * 28))
    y = tf.placeholder(tf.float32, shape=(None, 10))

    # x = n * (28 * 28)
    # weight = (28 * 28) * 10
    weight = tf.Variable(tf.zeros((28 * 28, 10)))
    bias = tf.Variable(tf.zeros((10,)))

    # h = n * 10
    h = tf.sigmoid(tf.matmul(x, weight) + bias)

    cost = tf.reduce_mean(-(tf.reduce_sum(y * tf.log(h), 1) + tf.reduce_sum((1. -y) * tf.log(1. -h), 1)))
    optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)

    init = tf.initialize_all_variables()
    batch_size = 100
    display_step = 1
    with tf.Session() as sess:
        sess.run(init)

        # Training cycle
        for epoch in range(10):
            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)
                # Fit training using batch data
                sess.run(optimizer, feed_dict={x: batch_xs, y: batch_ys})
                # Compute average loss
                avg_cost += sess.run(cost, feed_dict={x: batch_xs, y: batch_ys})/total_batch
            # Display logs per epoch step
            if epoch % display_step == 0:
                print( "Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f}".format(avg_cost))

        # Test model
        correct_prediction = tf.equal(tf.argmax(h, 1), tf.argmax(y, 1))
        # Calculate accuracy
        accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))
        print("Accuracy:", accuracy.eval({x: mnist.test.images, y: mnist.test.labels}))