Python object_detection.builders.box_predictor_builder.build_convolutional_keras_box_predictor() Examples
The following are 21
code examples of object_detection.builders.box_predictor_builder.build_convolutional_keras_box_predictor().
You can vote up the ones you like or vote down the ones you don't like,
and go to the original project or source file by following the links above each example.
You may also want to check out all available functions/classes of the module
object_detection.builders.box_predictor_builder
, or try the search function
.
.
Example #1
| Source File: convolutional_keras_box_predictor_test.py From vehicle_counting_tensorflow with MIT License | 5 votes |
|
def test_get_boxes_for_one_aspect_ratio_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[1],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND], axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 64, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 64, 1])
Example #2
| Source File: convolutional_keras_box_predictor_test.py From multilabel-image-classification-tensorflow with MIT License | 5 votes |
|
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
class_predictions_with_background = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, class_predictions_with_background)
(box_encodings,
class_predictions_with_background) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background.shape,
[4, 320, num_classes_without_background+1])
Example #3
| Source File: convolutional_keras_box_predictor_test.py From multilabel-image-classification-tensorflow with MIT License | 5 votes |
|
def test_get_boxes_for_one_aspect_ratio_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[1],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND], axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 64, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 64, 1])
Example #4
| Source File: convolutional_keras_box_predictor_test.py From multilabel-image-classification-tensorflow with MIT License | 5 votes |
|
def test_get_boxes_for_five_aspect_ratios_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 320, 1])
Example #5
| Source File: convolutional_keras_box_predictor_tf2_test.py From models with Apache License 2.0 | 5 votes |
|
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
def graph_fn(image_features):
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
class_predictions_with_background = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, class_predictions_with_background)
(box_encodings,
class_predictions_with_background) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background.shape,
[4, 320, num_classes_without_background+1])
Example #6
| Source File: convolutional_keras_box_predictor_tf2_test.py From models with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_one_aspect_ratio_per_location(self):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[1],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
def graph_fn(image_features):
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND], axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 64, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 64, 1])
Example #7
| Source File: convolutional_keras_box_predictor_tf2_test.py From models with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_five_aspect_ratios_per_location(self):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
def graph_fn(image_features):
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 320, 1])
Example #8
| Source File: convolutional_keras_box_predictor_test.py From g-tensorflow-models with Apache License 2.0 | 5 votes |
|
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
class_predictions_with_background = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, class_predictions_with_background)
(box_encodings,
class_predictions_with_background) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background.shape,
[4, 320, num_classes_without_background+1])
Example #9
| Source File: convolutional_keras_box_predictor_test.py From g-tensorflow-models with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_one_aspect_ratio_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[1],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND], axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 64, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 64, 1])
Example #10
| Source File: convolutional_keras_box_predictor_test.py From g-tensorflow-models with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_five_aspect_ratios_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 320, 1])
Example #11
| Source File: convolutional_keras_box_predictor_test.py From MAX-Object-Detector with Apache License 2.0 | 5 votes |
|
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
class_predictions_with_background = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, class_predictions_with_background)
(box_encodings,
class_predictions_with_background) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background.shape,
[4, 320, num_classes_without_background+1])
Example #12
| Source File: convolutional_keras_box_predictor_test.py From MAX-Object-Detector with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_one_aspect_ratio_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[1],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(box_predictions[
box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND], axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 64, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 64, 1])
Example #13
| Source File: convolutional_keras_box_predictor_test.py From MAX-Object-Detector with Apache License 2.0 | 5 votes |
|
def test_get_boxes_for_five_aspect_ratios_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 320, 1])
Example #14
| Source File: convolutional_keras_box_predictor_test.py From vehicle_counting_tensorflow with MIT License | 5 votes |
|
def test_get_multi_class_predictions_for_five_aspect_ratios_per_location(
self):
num_classes_without_background = 6
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=num_classes_without_background,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
class_predictions_with_background = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, class_predictions_with_background)
(box_encodings,
class_predictions_with_background) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(class_predictions_with_background.shape,
[4, 320, num_classes_without_background+1])
Example #15
| Source File: convolutional_keras_box_predictor_test.py From vehicle_counting_tensorflow with MIT License | 5 votes |
|
def test_get_boxes_for_five_aspect_ratios_per_location(self):
def graph_fn(image_features):
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return (box_encodings, objectness_predictions)
image_features = np.random.rand(4, 8, 8, 64).astype(np.float32)
(box_encodings, objectness_predictions) = self.execute(graph_fn,
[image_features])
self.assertAllEqual(box_encodings.shape, [4, 320, 1, 4])
self.assertAllEqual(objectness_predictions.shape, [4, 320, 1])
Example #16
| Source File: convolutional_keras_box_predictor_test.py From MAX-Object-Detector with Apache License 2.0 | 4 votes |
|
def test_get_predictions_with_feature_maps_of_dynamic_shape(
self):
image_features = tf.placeholder(dtype=tf.float32, shape=[4, None, None, 64])
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution*resolution*5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape,
objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings), tf.shape(objectness_predictions)],
feed_dict={image_features:
np.random.rand(4, resolution, resolution, 64)})
actual_variable_set = set(
[var.op.name for var in tf.trainable_variables()])
self.assertAllEqual(box_encodings_shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
self.assertEqual(conv_box_predictor._sorted_head_names,
['box_encodings', 'class_predictions_with_background'])
# TODO(kaftan): Remove conditional after CMLE moves to TF 1.10
Example #17
| Source File: convolutional_keras_box_predictor_test.py From g-tensorflow-models with Apache License 2.0 | 4 votes |
|
def test_get_predictions_with_feature_maps_of_dynamic_shape(
self):
image_features = tf.placeholder(dtype=tf.float32, shape=[4, None, None, 64])
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution*resolution*5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape,
objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings), tf.shape(objectness_predictions)],
feed_dict={image_features:
np.random.rand(4, resolution, resolution, 64)})
actual_variable_set = set(
[var.op.name for var in tf.trainable_variables()])
self.assertAllEqual(box_encodings_shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
self.assertEqual(conv_box_predictor._sorted_head_names,
['box_encodings', 'class_predictions_with_background'])
# TODO(kaftan): Remove conditional after CMLE moves to TF 1.10
Example #18
| Source File: convolutional_keras_box_predictor_tf2_test.py From models with Apache License 2.0 | 4 votes |
|
def test_get_predictions_with_feature_maps_of_dynamic_shape(
self):
tf.keras.backend.clear_session()
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
variables = []
def graph_fn(image_features):
box_predictions = conv_box_predictor([image_features])
variables.extend(list(conv_box_predictor.variables))
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return box_encodings, objectness_predictions
resolution = 32
expected_num_anchors = resolution*resolution*5
box_encodings, objectness_predictions = self.execute(
graph_fn, [np.random.rand(4, resolution, resolution, 64)])
actual_variable_set = set([var.name.split(':')[0] for var in variables])
self.assertAllEqual(box_encodings.shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions.shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
self.assertEqual(conv_box_predictor._sorted_head_names,
['box_encodings', 'class_predictions_with_background'])
Example #19
| Source File: convolutional_keras_box_predictor_tf2_test.py From models with Apache License 2.0 | 4 votes |
|
def test_use_depthwise_convolution(self):
tf.keras.backend.clear_session()
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=3,
box_code_size=4,
use_depthwise=True
))
variables = []
def graph_fn(image_features):
box_predictions = conv_box_predictor([image_features])
variables.extend(list(conv_box_predictor.variables))
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
return box_encodings, objectness_predictions
resolution = 32
expected_num_anchors = resolution*resolution*5
box_encodings, objectness_predictions = self.execute(
graph_fn, [np.random.rand(4, resolution, resolution, 64)])
actual_variable_set = set([var.name.split(':')[0] for var in variables])
self.assertAllEqual(box_encodings.shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions.shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor_depthwise/'
'bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor_depthwise/'
'depthwise_kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor_depthwise/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor_depthwise/'
'depthwise_kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
self.assertEqual(conv_box_predictor._sorted_head_names,
['box_encodings', 'class_predictions_with_background'])
Example #20
| Source File: convolutional_keras_box_predictor_test.py From vehicle_counting_tensorflow with MIT License | 4 votes |
|
def test_get_predictions_with_feature_maps_of_dynamic_shape(
self):
image_features = tf.placeholder(dtype=tf.float32, shape=[4, None, None, 64])
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution*resolution*5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape,
objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings), tf.shape(objectness_predictions)],
feed_dict={image_features:
np.random.rand(4, resolution, resolution, 64)})
actual_variable_set = set(
[var.op.name for var in tf.trainable_variables()])
self.assertAllEqual(box_encodings_shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
# TODO(kaftan): Remove conditional after CMLE moves to TF 1.10
Example #21
| Source File: convolutional_keras_box_predictor_test.py From multilabel-image-classification-tensorflow with MIT License | 4 votes |
|
def test_get_predictions_with_feature_maps_of_dynamic_shape(
self):
image_features = tf.placeholder(dtype=tf.float32, shape=[4, None, None, 64])
conv_box_predictor = (
box_predictor_builder.build_convolutional_keras_box_predictor(
is_training=False,
num_classes=0,
conv_hyperparams=self._build_conv_hyperparams(),
freeze_batchnorm=False,
inplace_batchnorm_update=False,
num_predictions_per_location_list=[5],
min_depth=0,
max_depth=32,
num_layers_before_predictor=1,
use_dropout=True,
dropout_keep_prob=0.8,
kernel_size=1,
box_code_size=4
))
box_predictions = conv_box_predictor([image_features])
box_encodings = tf.concat(
box_predictions[box_predictor.BOX_ENCODINGS], axis=1)
objectness_predictions = tf.concat(
box_predictions[box_predictor.CLASS_PREDICTIONS_WITH_BACKGROUND],
axis=1)
init_op = tf.global_variables_initializer()
resolution = 32
expected_num_anchors = resolution*resolution*5
with self.test_session() as sess:
sess.run(init_op)
(box_encodings_shape,
objectness_predictions_shape) = sess.run(
[tf.shape(box_encodings), tf.shape(objectness_predictions)],
feed_dict={image_features:
np.random.rand(4, resolution, resolution, 64)})
actual_variable_set = set(
[var.op.name for var in tf.trainable_variables()])
self.assertAllEqual(box_encodings_shape, [4, expected_num_anchors, 1, 4])
self.assertAllEqual(objectness_predictions_shape,
[4, expected_num_anchors, 1])
expected_variable_set = set([
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/bias',
'BoxPredictor/SharedConvolutions_0/Conv2d_0_1x1_32/kernel',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/bias',
'BoxPredictor/ConvolutionalBoxHead_0/BoxEncodingPredictor/kernel',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/bias',
'BoxPredictor/ConvolutionalClassHead_0/ClassPredictor/kernel'])
self.assertEqual(expected_variable_set, actual_variable_set)
# TODO(kaftan): Remove conditional after CMLE moves to TF 1.10
