mnist_cnn.py

# coding: utf-8

# # MNIST CNN

# In[1]:


import tensorflow as tf


# In[3]:


from tensorflow.examples.tutorials.mnist import input_data


# In[4]:


mnist = input_data.read_data_sets("MNIST_data/", one_hot=True)


# # Helper Functions
#

# ## INIT WEIGHTS

# In[21]:


def init_weights(shape):
    init_random_dist = tf.truncated_normal(shape, stddev=0.1)
    return tf.Variable(init_random_dist)


# ## INIT BIAS

# In[22]:


def init_bias(shape):
    init_bias_vals = tf.constant(0.1, shape=shape)
    return tf.Variable(init_bias_vals)


# ## CONV2D

# In[23]:


def conv2d(x, W):
    # x --> [batch,H,W,channels]
    # W --> [filter H, filter W, Channels In, Channels Out]
    return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding="SAME")


# ## MAX POOLING

# In[36]:


def max_pool_2x2(x):
    # x --> [batch,H,W,channels]
    return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding="SAME")


# ## CONVOLUTIONAL LAYER

# In[31]:


def convolutional_layer(input_x, shape):
    W = init_weights(shape)
    b = init_bias([shape[3]])
    return tf.nn.relu(conv2d(input_x, W)+b)


# ## FULLY CONNECTED LAYER

# In[26]:


def normal_full_layer(input_layer, size):
    input_size = int(input_layer.get_shape()[1])
    W = init_weights([input_size, size])
    b = init_bias([size])
    return tf.matmul(input_layer, W)+b


# In[33]:


# Placeholders
x = tf.placeholder(tf.float32, shape=[None, 784])
y_true = tf.placeholder(tf.float32, shape=[None, 10])
x_image = tf.reshape(x, [-1, 28, 28, 1])


# In[37]:


convo_1 = convolutional_layer(x_image, shape=[5, 5, 1, 32])
convo_1_pooling = max_pool_2x2(convo_1)


# In[38]:


convo_2 = convolutional_layer(convo_1_pooling, shape=[5, 5, 32, 64])
convo_2_pooling = max_pool_2x2(convo_2)


# In[40]:


# Why 7 by 7 image? Because we did 2 pooling layers, so (28/2)/2 = 7
convo_2_flat = tf.reshape(convo_2_pooling, [-1, 7*7*64])
full_layer_1 = tf.nn.relu(normal_full_layer(convo_2_flat, 1024))


# In[41]:


# Dropout
hold_prob = tf.placeholder(tf.float32)
full_one_dropout = tf.nn.dropout(full_layer_1, keep_prob=hold_prob)


# In[42]:


y_pred = normal_full_layer(full_one_dropout, 10)


# In[44]:


# Loss function
cross_entropy = tf.reduce_mean(
    tf.nn.softmax_cross_entropy_with_logits_v2(labels=y_true, logits=y_pred))


# In[45]:


# Optimizer
optimizer = tf.train.AdamOptimizer(learning_rate=0.001)
train = optimizer.minimize(cross_entropy)


# In[46]:


init = tf.global_variables_initializer()


# In[47]:


steps = 5001
with tf.Session() as sess:
    sess.run(init)
    for i in range(steps):
        batch_x, batch_y = mnist.train.next_batch(100)
        sess.run(train, feed_dict={x: batch_x,
                                   y_true: batch_y, hold_prob: 0.5})
        if i % 100 == 0:
            print("On Step: {}".format(i))
            print("Accuracy")
            matches = tf.equal(tf.argmax(y_pred, 1), tf.argmax(y_true, 1))
            acc = tf.reduce_mean(tf.cast(matches, tf.float32))
            print(sess.run(acc, feed_dict={
                  x: mnist.test.images, y_true: mnist.test.labels, hold_prob: 1.0}))