in research/pate_2017/deep_cnn.py [0:0]
def inference(images, dropout=False):
"""Build the CNN model.
Args:
images: Images returned from distorted_inputs() or inputs().
dropout: Boolean controlling whether to use dropout or not
Returns:
Logits
"""
if FLAGS.dataset == 'mnist':
first_conv_shape = [5, 5, 1, 64]
else:
first_conv_shape = [5, 5, 3, 64]
# conv1
with tf.variable_scope('conv1') as scope:
kernel = _variable_with_weight_decay('weights',
shape=first_conv_shape,
stddev=1e-4,
wd=0.0)
conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
biases = _variable_on_cpu('biases', [64], tf.constant_initializer(0.0))
bias = tf.nn.bias_add(conv, biases)
conv1 = tf.nn.relu(bias, name=scope.name)
if dropout:
conv1 = tf.nn.dropout(conv1, 0.3, seed=FLAGS.dropout_seed)
# pool1
pool1 = tf.nn.max_pool(conv1,
ksize=[1, 3, 3, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool1')
# norm1
norm1 = tf.nn.lrn(pool1,
4,
bias=1.0,
alpha=0.001 / 9.0,
beta=0.75,
name='norm1')
# conv2
with tf.variable_scope('conv2') as scope:
kernel = _variable_with_weight_decay('weights',
shape=[5, 5, 64, 128],
stddev=1e-4,
wd=0.0)
conv = tf.nn.conv2d(norm1, kernel, [1, 1, 1, 1], padding='SAME')
biases = _variable_on_cpu('biases', [128], tf.constant_initializer(0.1))
bias = tf.nn.bias_add(conv, biases)
conv2 = tf.nn.relu(bias, name=scope.name)
if dropout:
conv2 = tf.nn.dropout(conv2, 0.3, seed=FLAGS.dropout_seed)
# norm2
norm2 = tf.nn.lrn(conv2,
4,
bias=1.0,
alpha=0.001 / 9.0,
beta=0.75,
name='norm2')
# pool2
pool2 = tf.nn.max_pool(norm2,
ksize=[1, 3, 3, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool2')
# local3
with tf.variable_scope('local3') as scope:
# Move everything into depth so we can perform a single matrix multiply.
reshape = tf.reshape(pool2, [FLAGS.batch_size, -1])
dim = reshape.get_shape()[1].value
weights = _variable_with_weight_decay('weights',
shape=[dim, 384],
stddev=0.04,
wd=0.004)
biases = _variable_on_cpu('biases', [384], tf.constant_initializer(0.1))
local3 = tf.nn.relu(tf.matmul(reshape, weights) + biases, name=scope.name)
if dropout:
local3 = tf.nn.dropout(local3, 0.5, seed=FLAGS.dropout_seed)
# local4
with tf.variable_scope('local4') as scope:
weights = _variable_with_weight_decay('weights',
shape=[384, 192],
stddev=0.04,
wd=0.004)
biases = _variable_on_cpu('biases', [192], tf.constant_initializer(0.1))
local4 = tf.nn.relu(tf.matmul(local3, weights) + biases, name=scope.name)
if dropout:
local4 = tf.nn.dropout(local4, 0.5, seed=FLAGS.dropout_seed)
# compute logits
with tf.variable_scope('softmax_linear') as scope:
weights = _variable_with_weight_decay('weights',
[192, FLAGS.nb_labels],
stddev=1/192.0,
wd=0.0)
biases = _variable_on_cpu('biases',
[FLAGS.nb_labels],
tf.constant_initializer(0.0))
logits = tf.add(tf.matmul(local4, weights), biases, name=scope.name)
return logits