def testModelVariables(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
with self.test_session():
inputs = random_ops.random_uniform((batch_size, height, width, 3))
vgg.vgg_a(inputs, num_classes)
expected_names = [
'vgg_a/conv1/conv1_1/weights',
'vgg_a/conv1/conv1_1/biases',
'vgg_a/conv2/conv2_1/weights',
'vgg_a/conv2/conv2_1/biases',
'vgg_a/conv3/conv3_1/weights',
'vgg_a/conv3/conv3_1/biases',
'vgg_a/conv3/conv3_2/weights',
'vgg_a/conv3/conv3_2/biases',
'vgg_a/conv4/conv4_1/weights',
'vgg_a/conv4/conv4_1/biases',
'vgg_a/conv4/conv4_2/weights',
'vgg_a/conv4/conv4_2/biases',
'vgg_a/conv5/conv5_1/weights',
'vgg_a/conv5/conv5_1/biases',
'vgg_a/conv5/conv5_2/weights',
'vgg_a/conv5/conv5_2/biases',
'vgg_a/fc6/weights',
'vgg_a/fc6/biases',
'vgg_a/fc7/weights',
'vgg_a/fc7/biases',
'vgg_a/fc8/weights',
'vgg_a/fc8/biases',
]
model_variables = [
v.op.name for v in variables_lib.get_model_variables()
]
self.assertSetEqual(set(model_variables), set(expected_names))
def testForward(self):
batch_size = 1
height, width = 224, 224
with self.test_session() as sess:
inputs = random_ops.random_uniform((batch_size, height, width, 3))
logits, _ = vgg.vgg_a(inputs)
sess.run(variables.global_variables_initializer())
output = sess.run(logits)
self.assertTrue(output.any())
def testFullyConvolutional(self):
batch_size = 1
height, width = 256, 256
num_classes = 1000
with self.test_session():
inputs = random_ops.random_uniform((batch_size, height, width, 3))
logits, _ = vgg.vgg_a(inputs, num_classes, spatial_squeeze=False)
self.assertEquals(logits.op.name, 'vgg_a/fc8/BiasAdd')
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, 2, 2, num_classes])
def testBuild(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
with self.test_session():
inputs = random_ops.random_uniform((batch_size, height, width, 3))
logits, _ = vgg.vgg_a(inputs, num_classes)
self.assertEquals(logits.op.name, 'vgg_a/fc8/squeezed')
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, num_classes])
def testEvaluation(self):
batch_size = 2
height, width = 224, 224
num_classes = 1000
with self.test_session():
eval_inputs = tf.random_uniform((batch_size, height, width, 3))
logits, _ = vgg.vgg_a(eval_inputs, is_training=False)
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, num_classes])
predictions = tf.argmax(logits, 1)
self.assertListEqual(predictions.get_shape().as_list(),
[batch_size])
def testTrainEvalWithReuse(self):
train_batch_size = 2
eval_batch_size = 1
train_height, train_width = 224, 224
eval_height, eval_width = 256, 256
num_classes = 1000
with self.test_session():
train_inputs = random_ops.random_uniform(
(train_batch_size, train_height, train_width, 3))
logits, _ = vgg.vgg_a(train_inputs)
self.assertListEqual(logits.get_shape().as_list(),
[train_batch_size, num_classes])
variable_scope.get_variable_scope().reuse_variables()
eval_inputs = random_ops.random_uniform(
(eval_batch_size, eval_height, eval_width, 3))
logits, _ = vgg.vgg_a(eval_inputs,
is_training=False,
spatial_squeeze=False)
self.assertListEqual(logits.get_shape().as_list(),
[eval_batch_size, 2, 2, num_classes])
logits = math_ops.reduce_mean(logits, [1, 2])
predictions = math_ops.argmax(logits, 1)
self.assertEquals(predictions.get_shape().as_list(),
[eval_batch_size])
def testEndPoints(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
with self.test_session():
inputs = tf.random_uniform((batch_size, height, width, 3))
_, end_points = vgg.vgg_a(inputs, num_classes)
expected_names = [
'vgg_a/conv1/conv1_1', 'vgg_a/pool1', 'vgg_a/conv2/conv2_1',
'vgg_a/pool2', 'vgg_a/conv3/conv3_1', 'vgg_a/conv3/conv3_2',
'vgg_a/pool3', 'vgg_a/conv4/conv4_1', 'vgg_a/conv4/conv4_2',
'vgg_a/pool4', 'vgg_a/conv5/conv5_1', 'vgg_a/conv5/conv5_2',
'vgg_a/pool5', 'vgg_a/fc6', 'vgg_a/fc7', 'vgg_a/fc8'
]
self.assertSetEqual(set(end_points.keys()), set(expected_names))
def testEndPoints(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
for is_training in [True, False]:
with ops.Graph().as_default():
inputs = random_ops.random_uniform(
(batch_size, height, width, 3))
_, end_points = vgg.vgg_a(inputs,
num_classes,
is_training=is_training)
expected_names = [
'vgg_a/conv1/conv1_1', 'vgg_a/pool1',
'vgg_a/conv2/conv2_1', 'vgg_a/pool2',
'vgg_a/conv3/conv3_1', 'vgg_a/conv3/conv3_2',
'vgg_a/pool3', 'vgg_a/conv4/conv4_1',
'vgg_a/conv4/conv4_2', 'vgg_a/pool4',
'vgg_a/conv5/conv5_1', 'vgg_a/conv5/conv5_2',
'vgg_a/pool5', 'vgg_a/fc6', 'vgg_a/fc7', 'vgg_a/fc8'
]
self.assertSetEqual(set(end_points.keys()),
set(expected_names))
# net = slim.dropout(net, 0.5, scope='dropout6')
# net = slim.fully_connected(net, 4096, scope='fc7')
# if is_training:
# net = slim.dropout(net, 0.5, scope='dropout7')
# net = slim.fully_connected(net, 300, activation_fn=None, scope='fc8')
# return net
train_log_dir = './log'
if not tf.gfile.Exists(train_log_dir):
tf.gfile.MakeDirs(train_log_dir)
with tf.Graph().as_default():
# Set up the data loading:
images, labels = zjltf.get_tfrecord(BATCH_SIZE, isTrain=True)
# Define the model:
predictions, endp = vgg.vgg_a(images, is_training=True)
# Specify the loss function:
#print(predictions)
#print(endp)
slim.losses.softmax_cross_entropy(predictions, labels)
total_loss = slim.losses.get_total_loss()
tf.summary.scalar('losses/total_loss', total_loss)
# Specify the optimization scheme:
optimizer = tf.train.GradientDescentOptimizer(learning_rate=.001)
# create_train_op that ensures that when we evaluate it to get the loss,
# the update_ops are done and the gradient updates are computed.
train_tensor = slim.learning.create_train_op(total_loss, optimizer)