def mobilenetv2_head(inputs, is_training=True, stage='stage1'):
if stage == 'stage1':
with slim.arg_scope(
mobilenetv2_scope(is_training=is_training, trainable=True)):
flatten, _ = mobilenet_v2.mobilenet(
input_tensor=inputs,
num_classes=None,
is_training=False,
depth_multiplier=1.0,
scope='MobilenetV2',
conv_defs=V2_HEAD_DEF,
finegrain_classification_mode=False
) #有global average pooling输出1*1*1280
flatten = tf.squeeze(flatten, [1, 2])
else:
fc_flatten = slim.flatten(input)
net = slim.fully_connected(fc_flatten,
1024,
scope='fc_1_{}'.format(stage))
net = slim.dropout(net,
keep_prob=0.5,
is_training=is_training,
scope='dropout_{}'.format(stage))
flatten = slim.fully_connected(net,
1024,
scope='fc_2_{}'.format(stage))
return flatten
def mobilenetv2_head(inputs, is_training=True):
with slim.arg_scope(mobilenetv2_scope(is_training=is_training, trainable=True)):
net, _ = mobilenet_v2.mobilenet(input_tensor=inputs,
num_classes=None,
is_training=False,
depth_multiplier=1.0,
scope='MobilenetV2',
conv_defs=V2_HEAD_DEF,
finegrain_classification_mode=False)
net = tf.squeeze(net, [1, 2])
return net