def get_model(path, n_class):
from models.mobilenet_v2 import MobileNetV2
fullnet = MobileNetV2(num_classes=1000)
# fullnet not pruned
if not args.isfullnetpruned:
fullnet.load_state_dict(torch.load(args.load_path))
net = MobileNetV2_prescreen(fullnet)
del fullnet
# fullnet is pruned model: used for iterative pruning
else:
if args.isfullnetpruned:
net = MobileNetV2_prescreen(fullnet)
checkpoint = torch.load(path, map_location='cpu')
if 'state_dict' in checkpoint.keys():
checkpoint = checkpoint['state_dict']
from collections import OrderedDict
new_state_dict = OrderedDict()
for k, v in checkpoint.items():
if k[0:6] == 'module':
name = k[7:] # remove module.
else:
name = k
new_state_dict[name] = v
net.load_state_dict(new_state_dict)
del fullnet
if args.n_gpu > 1:
net = torch.nn.DataParallel(net, device_ids=gpu_list)
net = net.to(device)
else:
net = net.to(device)
return net
def model_fn(FLAGS, objective, optimizer, metrics):
input_layer = Input(shape=(FLAGS.input_size, FLAGS.input_size, 3))
base_model1 = Xception(weights="imagenet",
include_top=False,
pooling='avg',
input_shape=(FLAGS.input_size, FLAGS.input_size, 3),
classes=FLAGS.num_classes)
base_model2 = MobileNetV2(weights='imagenet',
include_top=False,
pooling='avg',
alpha=1.4,
input_shape=(FLAGS.input_size, FLAGS.input_size, 3),
classes=FLAGS.num_classes,
)
base_model3 = DenseNet201(weights="imagenet",
include_top=False,
pooling='avg',
input_shape=(FLAGS.input_size, FLAGS.input_size, 3),
classes=FLAGS.num_classes)
x1 = base_model1(inputs=input_layer)
x2 = base_model2(inputs=input_layer)
x3 = base_model3(inputs=input_layer)
# 值得修改的地方
x1 = Dense(FLAGS.num_classes, activation='relu')(x1)
x2 = Dense(FLAGS.num_classes, activation='relu')(x2)
x3 = Dense(FLAGS.num_classes, activation='relu')(x3)
x = concatenate([x1, x2, x3], axis=1)
predictions = Dense(FLAGS.num_classes, activation='softmax')(x)
model = Model(inputs=input_layer, outputs=predictions)
model.compile(loss=objective, optimizer=optimizer, metrics=metrics)
return model
def build_model(self):
print("build mobile_net_v2")
mobile_v2_model = MobileNetV2(input_shape=(224, 224, 3),
alpha=1.0,
include_top=False,
weights=None)
x = mobile_v2_model.output
x = GlobalAveragePooling2D()(x)
outputs = Dense(self.config.num_output,
use_bias=True,
name='Logits_77')(x)
self.model = Model(inputs=mobile_v2_model.inputs, outputs=outputs)
if hasattr(self.config, "weights"):
print("loading weights from", self.config.weights)
self.model.load_weights(self.config.weights, by_name=True)
self.model.compile(
loss=self.face_loss,
optimizer=self.config.optimizer,
metrics=[self.nme],
)
def get_model_and_checkpoint(model, dataset, checkpoint_path, n_gpu=1):
if model == 'mobilenet' and dataset == 'imagenet':
from models.mobilenet import MobileNet
net = MobileNet(n_class=1000)
elif model == 'mobilenetv2' and dataset == 'imagenet':
from models.mobilenet_v2 import MobileNetV2
net = MobileNetV2(n_class=1000)
else:
raise NotImplementedError
if torch.cuda.is_available():
sd = torch.load(checkpoint_path)
else:
sd = torch.load(checkpoint_path, map_location=torch.device('cpu'))
if 'state_dict' in sd: # a checkpoint but not a state_dict
sd = sd['state_dict']
sd = {k.replace('module.', ''): v for k, v in sd.items()}
net.load_state_dict(sd)
if torch.cuda.is_available():
net = net.cuda()
if torch.cuda.is_available() and n_gpu > 1:
net = torch.nn.DataParallel(net, range(n_gpu))
return net, deepcopy(
net.state_dict()) # deepcopy的时候会将对象的每一层复制一个单独的个体出来，与之前的个体完全没有关系了
def build_model(net='MobileNet',
input_shape=(224, 224, 3),
siamese_weights=None,
share=True):
if net == 'MobileNet':
base_model = MobileNet(include_top=False, input_shape=input_shape)
elif net == 'MobileNetV2':
base_model = MobileNetV2(include_top=False, input_shape=input_shape)
elif net == 'NASNetMobile':
base_model = NASNetMobile(include_top=False, input_shape=input_shape)
elif net == 'ResNet18':
base_model = ResNet18(include_top=False, input_shape=input_shape)
elif net == 'ResNet18V2':
base_model = ResNet18V2(include_top=False, input_shape=input_shape)
elif net == 'ResNet34':
base_model = ResNet34(include_top=False, input_shape=input_shape)
elif net == 'ResNet34V2':
base_model = ResNet34V2(include_top=False, input_shape=input_shape)
elif net == 'DenseNet21':
base_model = DenseNet(include_top=False,
blocks=[2, 2, 2, 2],
input_shape=input_shape,
name='a')
if share == False:
base_model_b = DenseNet(include_top=False,
blocks=[2, 2, 2, 2],
input_shape=input_shape,
name='b')
elif net == 'DenseNet69':
base_model = DenseNet(include_top=False,
blocks=[6, 8, 10, 8],
input_shape=input_shape)
if share == False:
base_model_b = DenseNet(include_top=False,
blocks=[6, 8, 10, 8],
input_shape=input_shape,
name='b')
elif net == 'DenseNet109':
base_model = DenseNet(include_top=False,
blocks=[6, 12, 18, 16],
input_shape=input_shape)
if share == False:
base_model_b = DenseNet(include_top=False,
blocks=[6, 12, 18, 16],
input_shape=input_shape,
name='b')
elif net == 'DenseShuffleV1_57_373':
base_model = DenseShuffleV1(include_top=False,
blocks=[6, 8, 12],
input_shape=input_shape,
num_shuffle_units=[3, 7, 3],
scale_factor=1.0,
bottleneck_ratio=1,
dropout_rate=0.5)
elif net == 'DenseShuffleV2_57_373':
base_model = DenseShuffleV2(include_top=False,
blocks=[6, 8, 12],
input_shape=input_shape,
num_shuffle_units=[3, 7, 3],
scale_factor=1.0,
bottleneck_ratio=1,
dropout_rate=0.5)
elif net == 'DenseShuffleV2_49_353':
base_model = DenseShuffleV2(include_top=False,
blocks=[6, 8, 8],
input_shape=input_shape,
num_shuffle_units=[3, 5, 3],
scale_factor=1.0,
bottleneck_ratio=1,
dropout_rate=0.5)
elif net == 'DenseShuffleV2_17_232':
base_model = DenseShuffleV2(include_top=False,
blocks=[2, 2, 2],
input_shape=input_shape,
num_shuffle_units=[2, 3, 2],
scale_factor=1.0,
bottleneck_ratio=1,
dropout_rate=0.5)
elif net == 'ShuffleNetV2':
base_model = ShuffleNetV2(include_top=False,
scale_factor=1.0,
pooling='avg',
input_shape=input_shape,
num_shuffle_units=[3, 7, 3],
bottleneck_ratio=1)
elif net == 'ShuffleNet':
base_model = ShuffleNet(include_top=False,
scale_factor=1.0,
pooling='avg',
input_shape=input_shape,
num_shuffle_units=[3, 7, 3],
bottleneck_ratio=1)
elif net == 'MobileNetV3Small':
base_model = MobileNetV3Small(include_top=False,
input_shape=input_shape)
elif net == 'SqueezeNet':
base_model = SqueezeNet(include_top=False, input_shape=input_shape)
else:
print('the network name you have entered is not supported yet')
sys.exit()
input_a = keras.layers.Input(shape=input_shape, name='input_a')
input_b = keras.layers.Input(shape=input_shape, name='input_b')
processed_a = base_model(input_a)
if share:
processed_b = base_model(input_b)
else:
processed_b = base_model_b(input_b)
#processed_a = keras.layers.Activation('sigmoid', name='sigmoid_a')(processed_a)
#processed_b = keras.layers.Activation('sigmoid', name='sigmoid_b')(processed_b)
normalize = keras.layers.Lambda(lambda x: K.l2_normalize(x, axis=-1),
name='normalize')
processed_a = normalize(processed_a)
processed_b = normalize(processed_b)
distance = keras.layers.Lambda(euclidean_distance,
output_shape=eucl_dist_output_shape,
name='dist')([processed_a, processed_b])
model = keras.models.Model([input_a, input_b], distance)
if siamese_weights is not None:
print('load siamses weights ....')
model.load_weights(siamese_weights)
print('hahahaha')
return model