model.py

import tensorflow as tf

image_size = 32
input_images = tf.placeholder(tf.float32,
                              shape=[None, image_size, image_size, 3], 
                              name="input_images")

# First CONV layer
kernel = tf.Variable(tf.truncated_normal([11, 11, 3, 96],
                                         dtype=tf.float32,
                                         stddev=1e-1),
                     name="conv1_weights")
conv = tf.nn.conv2d(input_images, kernel, [1, 4, 4, 1], padding="SAME")
bias = tf.Variable(tf.truncated_normal([96]))
conv_with_bias = tf.nn.bias_add(conv, bias)
conv1 = tf.nn.relu(conv_with_bias, name="conv1")

lrn1 = tf.nn.lrn(conv1,
                 alpha=1e-4,
                 beta=0.75,
                 depth_radius=2,
                 bias=2.0)

pooled_conv1 = tf.nn.max_pool(lrn1,
                              ksize=[1, 3, 3, 1],
                              strides=[1, 2, 2, 1],
                              padding="SAME",
                              name="pool1")

# Second CONV Layer 
kernel = tf.Variable(tf.truncated_normal([5, 5, 96, 256],
                                         dtype=tf.float32,
                                         stddev=1e-1),
                     name="conv2_weights")
conv = tf.nn.conv2d(pooled_conv1, kernel, [1, 4, 4, 1], padding="SAME")
bias = tf.Variable(tf.truncated_normal([256]), name="conv2_bias")
conv_with_bias = tf.nn.bias_add(conv, bias)
conv2 = tf.nn.relu(conv_with_bias, name="conv2")
lrn2 = tf.nn.lrn(conv2,
                 alpha=1e-4,
                 beta=0.75,
                 depth_radius=2,
                 bias=2.0)

pooled_conv2 = tf.nn.max_pool(lrn2,
                              ksize=[1, 3, 3, 1],
                              strides=[1, 2, 2, 1],
                              padding="SAME",
                              name="pool2")