def _resnet_small(self,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
include_root_block=True,
spatial_squeeze=True,
reuse=None,
scope='resnet_v1_small'):
"""A shallow and thin ResNet v1 for faster tests."""
block = resnet_v1.resnet_v1_block
blocks = [
block('block1', base_depth=1, num_units=3, stride=2),
block('block2', base_depth=2, num_units=3, stride=2),
block('block3', base_depth=4, num_units=3, stride=2),
block('block4', base_depth=8, num_units=2, stride=1),
]
return resnet_v1.resnet_v1(inputs,
blocks,
num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
include_root_block=include_root_block,
spatial_squeeze=spatial_squeeze,
reuse=reuse,
scope=scope)
def GetResnet50Subnetwork(self,
images,
is_training=False,
global_pool=False,
reuse=None):
"""Constructs resnet_v1_50 part of the DELF model.
Args:
images: A tensor of size [batch, height, width, channels].
is_training: Whether or not the model is in training mode.
global_pool: If True, perform global average pooling after feature
extraction. This may be useful for DELF's descriptor fine-tuning stage.
reuse: Whether or not the layer and its variables should be reused.
Returns:
net: A rank-4 tensor of size [batch, height_out, width_out, channels_out].
If global_pool is True, height_out = width_out = 1.
end_points: A set of activations for external use.
"""
block = resnet_v1.resnet_v1_block
blocks = [
block('block1', base_depth=64, num_units=3, stride=2),
block('block2', base_depth=128, num_units=4, stride=2),
block('block3', base_depth=256, num_units=6, stride=2),
]
if self._target_layer_type == 'resnet_v1_50/block4':
blocks.append(block('block4', base_depth=512, num_units=3, stride=1))
net, end_points = resnet_v1.resnet_v1(
images,
blocks,
is_training=is_training,
global_pool=global_pool,
reuse=reuse,
scope='resnet_v1_50')
return net, end_points
def _resnet_small(self,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
include_root_block=True,
reuse=None,
scope='resnet_v1_small'):
"""A shallow and thin ResNet v1 for faster tests."""
bottleneck = resnet_v1.bottleneck
blocks = [
resnet_utils.Block('block1', bottleneck,
[(4, 1, 1)] * 2 + [(4, 1, 2)]),
resnet_utils.Block('block2', bottleneck,
[(8, 2, 1)] * 2 + [(8, 2, 2)]),
resnet_utils.Block('block3', bottleneck,
[(16, 4, 1)] * 2 + [(16, 4, 2)]),
resnet_utils.Block('block4', bottleneck, [(32, 8, 1)] * 2)
]
return resnet_v1.resnet_v1(inputs,
blocks,
num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
include_root_block=include_root_block,
reuse=reuse,
scope=scope)
def _resnet_small(self,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
include_root_block=True,
spatial_squeeze=True,
reuse=None,
scope='resnet_v1_small'):
"""A shallow and thin ResNet v1 for faster tests."""
block = resnet_v1.resnet_v1_block
blocks = [
block('block1', base_depth=1, num_units=3, stride=2),
block('block2', base_depth=2, num_units=3, stride=2),
block('block3', base_depth=4, num_units=3, stride=2),
block('block4', base_depth=8, num_units=2, stride=1),
]
return resnet_v1.resnet_v1(inputs, blocks, num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
include_root_block=include_root_block,
spatial_squeeze=spatial_squeeze,
reuse=reuse,
scope=scope)
def _resnet_small(self,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
include_root_block=True,
reuse=None,
scope='resnet_v1_small'):
"""A shallow and thin ResNet v1 for faster tests."""
bottleneck = resnet_v1.bottleneck
blocks = [
resnet_utils.Block(
'block1', bottleneck, [(4, 1, 1)] * 2 + [(4, 1, 2)]),
resnet_utils.Block(
'block2', bottleneck, [(8, 2, 1)] * 2 + [(8, 2, 2)]),
resnet_utils.Block(
'block3', bottleneck, [(16, 4, 1)] * 2 + [(16, 4, 2)]),
resnet_utils.Block(
'block4', bottleneck, [(32, 8, 1)] * 2)]
return resnet_v1.resnet_v1(inputs, blocks, num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
include_root_block=include_root_block,
reuse=reuse,
scope=scope)
def GetResnet50Subnetwork(self,
images,
is_training=False,
global_pool=False,
reuse=None):
"""Constructs resnet_v1_50 part of the DELF model.
Args:
images: A tensor of size [batch, height, width, channels].
is_training: Whether or not the model is in training mode.
global_pool: If True, perform global average pooling after feature
extraction. This may be useful for DELF's descriptor fine-tuning stage.
reuse: Whether or not the layer and its variables should be reused.
Returns:
net: A rank-4 tensor of size [batch, height_out, width_out, channels_out].
If global_pool is True, height_out = width_out = 1.
end_points: A set of activations for external use.
"""
block = resnet_v1.resnet_v1_block
blocks = [
block('block1', base_depth=64, num_units=3, stride=2),
block('block2', base_depth=128, num_units=4, stride=2),
block('block3', base_depth=256, num_units=6, stride=2),
]
if self._target_layer_type == 'resnet_v1_50/block4':
blocks.append(block('block4', base_depth=512, num_units=3, stride=1))
net, end_points = resnet_v1.resnet_v1(
images,
blocks,
is_training=is_training,
global_pool=global_pool,
reuse=reuse,
scope='resnet_v1_50')
return net, end_points
def resnet_v1_50_C6(inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
spatial_squeeze=True,
reuse=None,
scope='resnet_v1_50',
initial_conv_name='conv1'):
"""Unlike the slim default we use a stride of 2 in the last block."""
blocks = [
resnet_v1_block('block1', base_depth=64, num_units=3, stride=1),
resnet_v1_block('block2', base_depth=128, num_units=4, stride=2),
resnet_v1_block('block3', base_depth=256, num_units=6, stride=2),
resnet_v1_block('block4', base_depth=512, num_units=3, stride=2),
resnet_v1_block('block5', base_depth=512, num_units=3, stride=2)
]
return resnet_v1.resnet_v1(inputs,
blocks,
num_classes,
is_training,
global_pool=global_pool,
output_stride=output_stride,
include_root_block=True,
spatial_squeeze=spatial_squeeze,
reuse=reuse,
scope=scope,
initial_conv_name=initial_conv_name)
