def testVariablesSetDevice(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
inputs = tf.random_uniform((batch_size, height, width, 3))
# Force all Variables to reside on the device.
with tf.variable_scope('on_cpu'), tf.device('/cpu:0'):
darknet.darknet_19(inputs, num_classes)
with tf.variable_scope('on_gpu'), tf.device('/gpu:0'):
darknet.darknet_19(inputs, num_classes)
for v in tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,
scope='on_cpu'):
self.assertDeviceEqual(v.device, '/cpu:0')
for v in tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,
scope='on_gpu'):
self.assertDeviceEqual(v.device, '/gpu:0')
def testTrainEvalWithReuse(self):
train_batch_size = 5
eval_batch_size = 2
height, width = 150, 150
num_classes = 1000
with self.test_session() as sess:
train_inputs = tf.random_uniform(
(train_batch_size, height, width, 3))
darknet.darknet_19(train_inputs, num_classes)
eval_inputs = tf.random_uniform(
(eval_batch_size, height, width, 3))
logits, _ = darknet.darknet_19(eval_inputs,
num_classes,
is_training=False,
reuse=True)
predictions = tf.argmax(logits, 1)
sess.run(tf.global_variables_initializer())
output = sess.run(predictions)
self.assertEquals(output.shape, (eval_batch_size, ))
def testGlobalPool(self):
batch_size = 2
height, width = 448, 448
num_classes = 1000
inputs = tf.random_uniform((batch_size, height, width, 3))
logits, end_points = darknet.darknet_19(inputs, num_classes)
self.assertTrue(logits.op.name.startswith('darknet_19/Logits'))
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, num_classes])
pre_pool = end_points[scope_prefix + 'Conv2D_18']
self.assertListEqual(pre_pool.get_shape().as_list(),
[batch_size, 14, 14, 1024])
def testEvaluation(self):
batch_size = 2
height, width = 224, 224
num_classes = 1000
with self.test_session() as sess:
eval_inputs = tf.random_uniform((batch_size, height, width, 3))
logits, _ = darknet.darknet_19(eval_inputs,
num_classes,
is_training=False)
predictions = tf.argmax(logits, 1)
sess.run(tf.global_variables_initializer())
output = sess.run(predictions)
self.assertEquals(output.shape, (batch_size, ))
def testBuildLogits(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
inputs = tf.random_uniform((batch_size, height, width, 3))
logits, end_points = darknet.darknet_19(inputs, num_classes)
predictions = end_points['Predictions']
self.assertTrue(logits.op.name.startswith('darknet_19/Logits'))
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, num_classes])
self.assertTrue(
predictions.op.name.startswith('darknet_19/Predictions'))
self.assertListEqual(predictions.get_shape().as_list(),
[batch_size, num_classes])
def testUnknownBatchSize(self):
batch_size = 1
height, width = 224, 224
num_classes = 1000
with self.test_session() as sess:
inputs = tf.placeholder(tf.float32, (None, height, width, 3))
logits, _ = darknet.darknet_19(inputs, num_classes)
self.assertTrue(logits.op.name.startswith('darknet_19/Logits'))
self.assertListEqual(logits.get_shape().as_list(),
[None, num_classes])
images = tf.random_uniform((batch_size, height, width, 3))
sess.run(tf.global_variables_initializer())
output = sess.run(logits, {inputs: images.eval()})
self.assertEquals(output.shape, (batch_size, num_classes))
def testAllEndPointsShapes(self):
batch_size = 5
height, width = 224, 224
num_classes = 1000
inputs = tf.random_uniform((batch_size, height, width, 3))
_, end_points = darknet.darknet_19(inputs, num_classes)
endpoints_shapes = {
scope_prefix + 'Conv2D_1': [batch_size, 224, 224, 32],
scope_prefix + 'MaxPool_1': [batch_size, 112, 112, 32],
#
scope_prefix + 'Conv2D_2': [batch_size, 112, 112, 64],
scope_prefix + 'MaxPool_2': [batch_size, 56, 56, 64],
#
scope_prefix + 'Conv2D_3': [batch_size, 56, 56, 128],
scope_prefix + 'Conv2D_4': [batch_size, 56, 56, 64],
scope_prefix + 'Conv2D_5': [batch_size, 56, 56, 128],
scope_prefix + 'MaxPool_3': [batch_size, 28, 28, 128],
#
scope_prefix + 'Conv2D_6': [batch_size, 28, 28, 256],
scope_prefix + 'Conv2D_7': [batch_size, 28, 28, 128],
scope_prefix + 'Conv2D_8': [batch_size, 28, 28, 256],
scope_prefix + 'MaxPool_4': [batch_size, 14, 14, 256],
#
scope_prefix + 'Conv2D_9': [batch_size, 14, 14, 512],
scope_prefix + 'Conv2D_10': [batch_size, 14, 14, 256],
scope_prefix + 'Conv2D_11': [batch_size, 14, 14, 512],
scope_prefix + 'Conv2D_12': [batch_size, 14, 14, 256],
scope_prefix + 'Conv2D_13': [batch_size, 14, 14, 512],
scope_prefix + 'MaxPool_5': [batch_size, 7, 7, 512],
#
scope_prefix + 'Conv2D_14': [batch_size, 7, 7, 1024],
scope_prefix + 'Conv2D_15': [batch_size, 7, 7, 512],
scope_prefix + 'Conv2D_16': [batch_size, 7, 7, 1024],
scope_prefix + 'Conv2D_17': [batch_size, 7, 7, 512],
scope_prefix + 'Conv2D_18': [batch_size, 7, 7, 1024],
#
'Conv2D_19': [batch_size, 7, 7, num_classes],
# Logits and predictions
'global_pool': [batch_size, 1, 1, num_classes],
'Logits': [batch_size, num_classes],
'Predictions': [batch_size, num_classes]
}
self.assertItemsEqual(endpoints_shapes.keys(), end_points.keys())
for endpoint_name in endpoints_shapes:
expected_shape = endpoints_shapes[endpoint_name]
self.assertTrue(endpoint_name in end_points)
self.assertListEqual(
end_points[endpoint_name].get_shape().as_list(),
expected_shape)
def testGlobalPoolUnknownImageShape(self):
batch_size = 2
height, width = 448, 448
num_classes = 1000
with self.test_session() as sess:
inputs = tf.placeholder(tf.float32, (batch_size, None, None, 3))
logits, end_points = darknet.darknet_19(inputs, num_classes)
self.assertTrue(logits.op.name.startswith('darknet_19/Logits'))
self.assertListEqual(logits.get_shape().as_list(),
[batch_size, num_classes])
pre_pool = end_points[scope_prefix + 'Conv2D_18']
images = tf.random_uniform((batch_size, height, width, 3))
sess.run(tf.global_variables_initializer())
logits_out, pre_pool_out = sess.run([logits, pre_pool],
{inputs: images.eval()})
self.assertTupleEqual(logits_out.shape, (batch_size, num_classes))
self.assertTupleEqual(pre_pool_out.shape,
(batch_size, 14, 14, 1024))