def mobilenet_ssd(image, L2_reg, is_training=True):
assert 'MobilenetV2' in cfg.MODEL.net_structure
if cfg.TRAIN.lock_basenet_bn:
arg_scope = training_scope(weight_decay=L2_reg, is_training=False)
else:
arg_scope = training_scope(weight_decay=L2_reg,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
# _,endpoint = mobilenet_v2_050(image,is_training=is_training,base_only=True,finegrain_classification_mode=False)
_, endpoint = mobilenet_v2_140(image,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
for k, v in endpoint.items():
print('mobile backbone output:', k, v)
mobilebet_fms = [
endpoint['layer_5/expansion_output'],
endpoint['layer_8/expansion_output'],
endpoint['layer_15/expansion_output'], endpoint['layer_18/output']
]
print('mobile backbone output:', mobilebet_fms)
with slim.arg_scope(
resnet_arg_scope(weight_decay=L2_reg, bn_is_training=is_training)):
net = slim.conv2d(mobilebet_fms[-1],
128, [1, 1],
stride=1,
activation_fn=tf.nn.relu,
scope='extra_conv_1_1')
net = slim.conv2d(net,
256, [3, 3],
stride=2,
activation_fn=tf.nn.relu,
scope='extra_conv_1_2')
mobilebet_fms.append(net)
net = slim.conv2d(net,
128, [1, 1],
stride=1,
activation_fn=tf.nn.relu,
scope='extra_conv_2_1')
net = slim.conv2d(net,
256, [3, 3],
stride=2,
activation_fn=tf.nn.relu,
scope='extra_conv_2_2')
mobilebet_fms.append(net)
print('extra backbone output:', mobilebet_fms)
if cfg.MODEL.fpn:
enhanced_fms = create_fem_net(mobilebet_fms, L2_reg, is_training)
else:
enhanced_fms = None
return mobilebet_fms, enhanced_fms
def mobilenetv3_large(image, is_training=True):
arg_scope = training_scope(weight_decay=cfg.TRAIN.weight_decay_factor,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
_, endpoints = mobilnet_v3.large(image,
depth_multiplier=0.75,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
# end_feature=slim.separable_conv2d(endpoints['layer_16/output'],
# 512,
# [3, 3],
# stride=1,
# scope='mbntev3_extra_conv')
# for k,v in endpoints.items():
# print('mobile backbone output:',k,v)
#
mobilebet_fms = [
endpoints['layer_5/expansion_output'],
endpoints['layer_8/expansion_output'],
endpoints['layer_14/expansion_output'], _
]
return mobilebet_fms
def mobilenetv3_ssd(image, is_training=True):
arg_scope = training_scope(weight_decay=cfg.TRAIN.weight_decay_factor,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
_, endpoint = mobilnet_v3.small_minimalistic(
image,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
for k, v in endpoint.items():
print('mobile backbone output:', k, v)
mobilebet_fms = [
endpoint['layer_4/output'], endpoint['layer_8/output'],
endpoint['layer_12/output']
]
if cfg.MODEL.fpn:
mobilebet_fms = create_fpn_net(mobilebet_fms,
dims_list=cfg.MODEL.fpn_dims)
return mobilebet_fms
def mobilenetv3_small_minimalistic(image, is_training=True):
arg_scope = training_scope(weight_decay=cfg.TRAIN.weight_decay_factor,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
final_feature, endpoints = mobilnet_v3.small_minimalistic(
image,
depth_multiplier=1.0,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
x = tf.reduce_mean(final_feature, axis=[1, 2], keep_dims=True)
x = slim.dropout(x, 0.5)
x = slim.conv2d(x,
num_outputs=cfg.MODEL.cls,
kernel_size=[1, 1],
stride=1,
padding='VALID',
activation_fn=None,
normalizer_fn=None,
scope='classifier/0')
x = tf.squeeze(x, axis=1)
x = tf.squeeze(x, axis=1)
x = tf.identity(x, name='cls_output')
return x
def mobilenetv3_large_detection(image, is_training=True):
arg_scope = training_scope(weight_decay=cfg.TRAIN.weight_decay_factor,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
_, endpoints = mobilnet_v3.large(image,
depth_multiplier=0.75,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
for k, v in endpoints.items():
print('mobile backbone output:', k, v)
extern_conv = slim.conv2d(_,
480, [1, 1],
stride=1,
padding='SAME',
activation_fn=hard_swish,
scope='extern1')
print(extern_conv)
mobilebet_fms = [
endpoints['layer_5/expansion_output'],
endpoints['layer_7/expansion_output'],
endpoints['layer_13/output'], extern_conv
]
return mobilebet_fms
def mobilenetv3_small_minimalistic(image, is_training=True):
arg_scope = training_scope(weight_decay=cfg.TRAIN.weight_decay_factor,
is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
if cfg.DATA.channel == 1:
if cfg.MODEL.global_stride == 8:
stride = 2
else:
stride = 1
image = slim.separable_conv2d(image,
3, [3, 3],
stride=stride,
padding='SAME',
scope='preconv')
final_feature, endpoints = mobilnet_v3.small_minimalistic(
image,
depth_multiplier=1.0,
is_training=is_training,
base_only=True,
finegrain_classification_mode=False)
extern_conv = slim.separable_conv2d(final_feature,
128, [3, 3],
stride=2,
padding='SAME',
scope='extern1')
extern_conv = slim.separable_conv2d(extern_conv,
96, [3, 3],
padding='SAME',
scope='extern2')
extern_conv = slim.separable_conv2d(extern_conv,
128, [3, 3],
padding='SAME',
scope='extern3')
for k, v in endpoints.items():
print('mobile backbone output:', k, v)
mobilebet_fms = [
endpoints['layer_3/expansion_output'],
endpoints['layer_5/expansion_output'],
endpoints['layer_9/expansion_output'],
#final_feature,
extern_conv
]
return mobilebet_fms
def training_scope(**kwargs):
"""Defines MobilenetV2 training scope.
Usage:
with tf.contrib.slim.arg_scope(mobilenet_v2.training_scope()):
logits, endpoints = mobilenet_v2.mobilenet(input_tensor)
with slim.
Args:
**kwargs: Passed to mobilenet.training_scope. The following parameters
are supported:
weight_decay- The weight decay to use for regularizing the model.
stddev- Standard deviation for initialization, if negative uses xavier.
dropout_keep_prob- dropout keep probability
bn_decay- decay for the batch norm moving averages.
Returns:
An `arg_scope` to use for the mobilenet v2 model.
"""
return lib.training_scope(**kwargs)
def mobilenet_ssd(image,L2_reg,is_training=True):
arg_scope = training_scope(weight_decay=L2_reg, is_training=is_training)
with tf.contrib.slim.arg_scope(arg_scope):
_,endpoint = mobilenet_v2_035(image,is_training=is_training,base_only=True,finegrain_classification_mode=False)
for k,v in endpoint.items():
print('mobile backbone output:',k,v)
mobilebet_fms=[
endpoint['layer_8/expansion_output'],
endpoint['layer_15/expansion_output'],
endpoint['layer_18/output']]
if cfg.MODEL.fpn:
mobilebet_fms=create_fpn_net(mobilebet_fms,dims_list=cfg.MODEL.fpn_dims)
return mobilebet_fms