def get_network_byname(net_name,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
spatial_squeeze=True):
if net_name == 'resnet_v1_50':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
resnet_v1.resnet_arg_scope(weight_decay=FLAGS.weight_decay)):
logits, end_points = resnet_v1.resnet_v1_50(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
if net_name == 'resnet_v1_101':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
resnet_v1.resnet_arg_scope(weight_decay=FLAGS.weight_decay)):
logits, end_points = resnet_v1.resnet_v1_101(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
def get_network_byname(net_name,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
spatial_squeeze=True):
if net_name not in ['resnet_v1_50', 'mobilenet_224', 'inception_resnet', 'vgg16', 'resnet_v1_101']:
raise ValueError('''not include network: {}, net_name must in [resnet_v1_50, mobilenet_224,
inception_resnet, vgg16, resnet_v1_101]
'''.format(net_name))
if net_name == 'resnet_v1_50':
with slim.arg_scope(resnet_v1.resnet_arg_scope(weight_decay=cfgs.WEIGHT_DECAY[net_name])):
logits, end_points = resnet_v1.resnet_v1_50(inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze
)
return logits, end_points
if net_name == 'resnet_v1_101':
with slim.arg_scope(resnet_v1.resnet_arg_scope(weight_decay=cfgs.WEIGHT_DECAY[net_name])):
logits, end_points = resnet_v1.resnet_v1_101(inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze
)
return logits, end_points
def get_network_byname(net_name,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
spatial_squeeze=True):
if net_name not in [
'resnet_v1_50', 'mobilenet_224', 'inception_resnet', 'vgg16',
'resnet_v1_101'
]:
raise ValueError(
'''not include network: {}, net_name must in [resnet_v1_50, mobilenet_224,
inception_resnet, vgg16, resnet_v1_101]
'''.format(net_name))
if net_name == 'resnet_v1_50':
with slim.arg_scope(
resnet_v1.resnet_arg_scope(
weight_decay=cfgs.WEIGHT_DECAY[net_name])):
logits, end_points = resnet_v1.resnet_v1_50(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
if net_name == 'resnet_v1_101':
with slim.arg_scope(
resnet_v1.resnet_arg_scope(
weight_decay=cfgs.WEIGHT_DECAY[net_name])):
logits, end_points = resnet_v1.resnet_v1_101(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
def get_network_byname(net_name,
inputs,
num_classes=None,
is_training=True,
global_pool=True,
output_stride=None,
spatial_squeeze=True):
if net_name == 'resnet_v1_50':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
resnet_v1.resnet_arg_scope(weight_decay=FLAGS.weight_decay)):
logits, end_points = resnet_v1.resnet_v1_50(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
if net_name == 'resnet_v1_101':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
resnet_v1.resnet_arg_scope(weight_decay=FLAGS.weight_decay)):
logits, end_points = resnet_v1.resnet_v1_101(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
global_pool=global_pool,
output_stride=output_stride,
spatial_squeeze=spatial_squeeze)
return logits, end_points
if net_name == 'pvanet':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
pvanet.pvanet_scope(
is_training=is_training,
weights_initializer=slim.xavier_initializer(),
batch_norm_param_initializer=None,
beta_initializer=tf.zeros_initializer(),
gamma_initializer=tf.ones_initializer(),
weight_decay=0.99)):
logits, end_points = pvanet.pvanet(net=inputs,
include_last_bn_relu=True)
return logits, end_points
if net_name == 'vgg_16':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
vgg.vgg_arg_scope(weight_decay=FLAGS.weight_decay)):
logits, end_points = vgg.vgg_16(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
dropout_keep_prob=0.5,
spatial_squeeze=spatial_squeeze,
)
return logits, end_points
# if net_name == 'inception_resnet_v2':
# FLAGS = get_flags_byname(net_name)
# with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(weight_decay=FLAGS.weight_decay)):
# logits, end_points = inception_resnet_v2.inception_resnet_v2(inputs=inputs,
# num_classes=num_classes,
# is_training=is_training,
# dropout_keep_prob=0.8,
# )
# return logits, end_points
if net_name == 'inception_resnet':
FLAGS = get_flags_byname(net_name)
arg_sc = inception_resnet_v2.inception_resnet_v2_arg_scope(
weight_decay=FLAGS.weight_decay)
with slim.arg_scope(arg_sc):
logits, end_points = inception_resnet_v2.inception_resnet_v2(
inputs=inputs,
num_classes=num_classes,
is_training=is_training)
return logits, end_points
if net_name == 'inception_v4':
FLAGS = get_flags_byname(net_name)
arg_sc = inception_v4.inception_v4_arg_scope(
weight_decay=FLAGS.weight_decay)
with slim.arg_scope(arg_sc):
logits, end_points = inception_v4.inception_v4(
inputs=inputs,
num_classes=num_classes,
is_training=is_training)
return logits, end_points
if net_name == 'mobilenet_224':
FLAGS = get_flags_byname(net_name)
with slim.arg_scope(
mobilenet_v1.mobilenet_v1_arg_scope(
weight_decay=FLAGS.weight_decay)):
logits, end_points = mobilenet_v1.mobilenet_v1(
inputs=inputs,
num_classes=num_classes,
is_training=is_training,
spatial_squeeze=spatial_squeeze)
return logits, end_points