def VioNet_densenet(config, home_path):
device = config.device
ft_begin_idx = config.ft_begin_idx
sample_size = config.sample_size[0]
sample_duration = config.sample_duration
model = densenet121(num_classes=2,
sample_size=sample_size,
sample_duration=sample_duration).to(device)
# state_dict = torch.load(g_path +'/VioNet/'+ 'weights/DenseNet_Kinetics.pth')
state_dict = torch.load(
os.path.join(home_path, VIONET_WEIGHTS, 'DenseNet_Kinetics.pth'))
model.load_state_dict(state_dict)
params = dn.get_fine_tuning_params(model, ft_begin_idx)
return model, params
def CreatNet(opt):
name = opt.model_name
label_num = opt.label
if name =='lstm':
net = lstm.lstm(100,27,num_classes=label_num)
elif name == 'cnn_1d':
net = cnn_1d.cnn(opt.input_nc,num_classes=label_num)
elif name == 'resnet18_1d':
net = resnet_1d.resnet18()
net.conv1 = nn.Conv1d(opt.input_nc, 64, 7, 2, 3, bias=False)
net.fc = nn.Linear(512, label_num)
elif name == 'resnet34_1d':
net = resnet_1d.resnet34()
net.conv1 = nn.Conv1d(opt.input_nc, 64, 7, 2, 3, bias=False)
net.fc = nn.Linear(512, label_num)
elif name == 'multi_scale_resnet_1d':
net = multi_scale_resnet_1d.Multi_Scale_ResNet(inchannel=opt.input_nc, num_classes=label_num)
elif name == 'micro_multi_scale_resnet_1d':
net = micro_multi_scale_resnet_1d.Multi_Scale_ResNet(inchannel=opt.input_nc, num_classes=label_num)
elif name == 'multi_scale_resnet':
net = multi_scale_resnet.Multi_Scale_ResNet(inchannel=opt.input_nc, num_classes=label_num)
elif name == 'dfcnn':
net = dfcnn.dfcnn(num_classes = label_num)
elif name in ['resnet101','resnet50','resnet18']:
if name =='resnet101':
net = resnet.resnet101(pretrained=False)
net.fc = nn.Linear(2048, label_num)
elif name =='resnet50':
net = resnet.resnet50(pretrained=False)
net.fc = nn.Linear(2048, label_num)
elif name =='resnet18':
net = resnet.resnet18(pretrained=False)
net.fc = nn.Linear(512, label_num)
net.conv1 = nn.Conv2d(opt.input_nc, 64, 7, 2, 3, bias=False)
elif 'densenet' in name:
if name =='densenet121':
net = densenet.densenet121(pretrained=False,num_classes=label_num)
elif name == 'densenet201':
net = densenet.densenet201(pretrained=False,num_classes=label_num)
elif name =='squeezenet':
net = squeezenet.squeezenet1_1(pretrained=False,num_classes=label_num,inchannel = 1)
return net
def save_checkpoint():
"""
Save checkpoint with correct input and output names
Inputs: None
Ouutputs: None
"""
config = XRAYconfig()
images = tf.placeholder(tf.float32,
shape=[None, 224, 224, 3],
name='inputs')
labels = tf.placeholder(tf.int32, shape=[None], name='labels')
with tf.variable_scope('densenet121') as densenet_scope:
processed_image = imagenet_preprocessing(images)
with slim.arg_scope(
densenet.densenet_arg_scope(
weight_decay=config.l2_reg,
batch_norm_decay=config.batch_norm_decay,
batch_norm_epsilon=config.batch_norm_epsilon)):
target_logits, _ = densenet.densenet121(
inputs=processed_image,
num_classes=config.output_shape,
is_training=False,
scope=densenet_scope)
oh_enc = tf.one_hot(labels, config.output_shape)
masked_logits = tf.multiply(target_logits,
oh_enc,
name='masked_logits')
prob = tf.nn.softmax(target_logits, name='probability')
saver = tf.train.Saver()
restorer = tf.train.Saver()
sess = tf.Session()
sess.run(
tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer()))
restorer.restore(sess, config.restore_checkpoint)
save_path = saver.save(sess, "./model_to_freeze.ckpt")
print("Model saved in path: %s" % save_path)
sess.close()
def CreatNet(name):
if name == 'lstm':
net = lstm.lstm(100, 27, num_classes=5)
elif name == 'cnn_1d':
net = cnn_1d.cnn(1, num_classes=5)
elif name == 'resnet18_1d':
net = resnet_1d.resnet18()
net.conv1 = nn.Conv1d(1, 64, 7, 2, 3, bias=False)
net.fc = nn.Linear(512, 5)
elif name == 'multi_scale_resnet_1d':
net = multi_scale_resnet_1d.Multi_Scale_ResNet(inchannel=1,
num_classes=5)
elif name == 'multi_scale_resnet':
net = multi_scale_resnet.Multi_Scale_ResNet(inchannel=1, num_classes=5)
elif name == 'dfcnn':
net = dfcnn.dfcnn(num_classes=5)
elif name in ['resnet101', 'resnet50', 'resnet18']:
if name == 'resnet101':
net = resnet.resnet101(pretrained=False)
net.fc = nn.Linear(2048, 5)
elif name == 'resnet50':
net = resnet.resnet50(pretrained=False)
net.fc = nn.Linear(2048, 5)
elif name == 'resnet18':
net = resnet.resnet18(pretrained=False)
net.fc = nn.Linear(512, 5)
net.conv1 = nn.Conv2d(1, 64, 7, 2, 3, bias=False)
elif 'densenet' in name:
if name == 'densenet121':
net = densenet.densenet121(pretrained=False, num_classes=5)
elif name == 'densenet201':
net = densenet.densenet201(pretrained=False, num_classes=5)
elif name == 'squeezenet':
net = squeezenet.squeezenet1_1(pretrained=False,
num_classes=5,
inchannel=1)
return net
def generate_model(opt):
assert opt.mode in ['score', 'feature']
if opt.mode == 'score':
last_fc = True
elif opt.mode == 'feature':
last_fc = False
assert opt.model_name in ['resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet']
if opt.model_name == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
polices = resnet.get_fine_tuning_parameters(model, opt.ft_begin_index)
elif opt.model_name == 'wideresnet':
assert opt.model_depth in [50]
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, k=opt.wide_resnet_k,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'resnext':
assert opt.model_depth in [50, 101, 152]
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
return model, polices
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
args = {
"num_classes": opt.n_classes,
"shortcut_type": opt.resnet_shortcut,
"sample_size": opt.sample_size,
"sample_duration": opt.sample_duration
}
if opt.model_depth == 10:
model = resnet.resnet10(**args)
elif opt.model_depth == 18:
model = resnet.resnet18(**args)
elif opt.model_depth == 34:
model = resnet.resnet34(**args)
elif opt.model_depth == 50:
model = resnet.resnet50(**args)
elif opt.model_depth == 101:
model = resnet.resnet101(**args)
elif opt.model_depth == 152:
model = resnet.resnet152(**args)
elif opt.model_depth == 200:
model = resnet.resnet200(**args)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
args = {
"num_classes": opt.n_classes,
"shortcut_type": opt.resnet_shortcut,
"cardinality": opt.resnext_cardinality,
"sample_size": opt.sample_size,
"sample_duration": opt.sample_duration
}
if opt.model_depth == 50:
model = resnext.resnet50(**args)
elif opt.model_depth == 101:
model = resnext.resnet101(**args)
elif opt.model_depth == 152:
model = resnext.resnet152(**args)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
args = {
"num_classes": opt.n_classes,
"shortcut_type": opt.resnet_shortcut,
"sample_size": opt.sample_size,
"sample_duration": opt.sample_duration
}
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(**args)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(**args)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(**args)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(**args)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(**args)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(**args)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
args = {
"num_classes": opt.n_classes,
"sample_size": opt.sample_size,
"sample_duration": opt.sample_duration
}
if opt.model_depth == 121:
model = densenet.densenet121(**args)
elif opt.model_depth == 169:
model = densenet.densenet169(**args)
elif opt.model_depth == 201:
model = densenet.densenet201(**args)
elif opt.model_depth == 264:
model = densenet.densenet264(**args)
if opt.no_cuda:
device = 'cpu'
else:
device = 'cuda'
model = model.to(device)
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path, map_location=device)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features, opt.n_finetune_classes)
model.module.classifier = model.module.classifier.to(device)
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.to(device)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
# -*- coding:utf-8 -*-
# @time :2019.09.07
# @IDE : pycharm
# @author :lxztju
# @github : https://github.com/lxztju
import torch
from models.densenet import densenet121
import cfg
from load_data import val_dataloader, val_datasets
##定义模型的框架
model = densenet121(num_classes=cfg.NUM_CLASSES)
print(model)
##将模型放置在gpu上运行
if torch.cuda.is_available():
model.cuda()
###读取网络模型的键值对
trained_model = cfg.TRAINED_MODEL
state_dict = torch.load(trained_model)
# create new OrderedDict that does not contain `module.`
##由于之前的模型是在多gpu上训练的，因而保存的模型参数，键前边有‘module’，需要去掉，和训练模型一样构建新的字典
from collections import OrderedDict
new_state_dict = OrderedDict()
for k, v in state_dict.items():
head = k[:7]
if head == 'module.':
name = k[7:] # remove `module.`
else:
def test(file, class_names, data_dir, results_dir):
import platform
print(platform.platform())
import sys
print('Python ', sys.version)
import pydicom
print('pydicom ', pydicom.__version__)
# Sets device to GPU if available, else CPU
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') # pylint: disable=no-member
print('Using device:', device)
# Additional about GPU
if device.type == 'cuda':
print(torch.cuda.get_device_name(0))
print('Memory Usage:')
print('Allocated:', round(torch.cuda.memory_allocated(0) / 1024**3, 1),
'GB')
print('Cached: ', round(torch.cuda.memory_cached(0) / 1024**3, 1),
'GB')
# Optimiza la corrida
cudnn.benchmark = True
# Conjunto de datos con las transformaciones especificadas anteriormente
adni_dataset = NumpyADNI_Dataset(data_dir=data_dir)
# Entrenamiento y prueba
train_size = int(0.75 * len(adni_dataset))
test_size = len(adni_dataset) - train_size
_, test_dataset = torch.utils.data.random_split(adni_dataset,
[train_size, test_size])
test_loader = torch.utils.data.DataLoader(test_dataset,
batch_size=1,
num_workers=4)
print('%d MRI images in testing loader...' % (test_size))
# Inicializa, carga y corre el modelo
model = densenet121(channels=1,
num_classes=len(class_names),
drop_rate=0.7).cuda()
model = torch.nn.DataParallel(model).to(device)
model.load_state_dict(torch.load(results_dir + '/' + file))
model.eval()
test = []
predicted = []
with torch.no_grad():
for data in test_loader:
# get the inputs; data is a list of [inputs, labels]
inputs, labels = data
labels = labels.to(device)
_, label = torch.max(labels, 1) # pylint: disable=no-member
test.append(label)
outputs = model(inputs)
_, predicted_value = torch.max(outputs.data, 1) # pylint: disable=no-member
predicted.append(predicted_value)
test = [x.item() for x in test]
predicted = [x.item() for x in predicted]
# Imprime estadísticas y gráficos
print_info_and_plots(test, predicted, class_names)
def get_network(args):
""" return given network
"""
if args.net == 'vgg11':
from models.vgg import vgg11_bn
net = vgg11_bn()
elif args.net == 'vgg13':
from models.vgg import vgg13_bn
net = vgg13_bn()
elif args.net == 'vgg16':
from models.vgg import vgg16_bn
net = vgg16_bn()
elif args.net == 'vgg19':
from models.vgg import vgg19_bn
net = vgg19_bn()
elif args.net == 'googlenet':
from models.googLeNet import GoogLeNet
net = GoogLeNet()
elif args.net == 'inceptionv3':
from models.inceptionv3 import Inceptionv3
net = Inceptionv3()
elif args.net == 'resnet18':
from models.resnet import resnet18
net = resnet18()
elif args.net == 'resnet34':
from models.resnet import resnet34
net = resnet34()
elif args.net == 'resnet50':
from models.resnet import resnet50
net = resnet50()
elif args.net == 'resnet101':
from models.resnet import resnet101
net = resnet101()
elif args.net == 'resnet152':
from models.resnet import resnet152
net = resnet152()
elif args.net == 'wrn':
from models.wideresnet import wideresnet
net = wideresnet()
elif args.net == 'densenet121':
from models.densenet import densenet121
net = densenet121()
elif args.net == 'densenet161':
from models.densenet import densenet161
net = densenet161()
elif args.net == 'densenet169':
from models.densenet import densenet169
net = densenet169()
elif args.net == 'densenet201':
from models.densenet import densenet201
net = densenet201()
else:
print('the network name you have entered is not supported yet')
sys.exit()
if args.gpu:
print("use gpu")
net = net.cuda()
return net
def generate_model(opt):
assert opt.mode in ['score', 'feature']
if opt.mode == 'score':
last_fc = True
elif opt.mode == 'feature':
last_fc = False
assert opt.model_name in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model_name == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'wideresnet':
assert opt.model_depth in [50]
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'resnext':
assert opt.model_depth in [50, 101, 152]
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_name == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
last_fc=last_fc)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
return model
def generate_C3D_model(opt):
assert opt.mode in ['score', 'feature']
if opt.mode == 'score':
last_fc = True
elif opt.mode == 'feature':
last_fc = False
assert opt.c3d_model_name in ['resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet']
if opt.c3d_model_name == 'resnet':
assert opt.c3d_model_depth in [10, 18, 34, 50, 101, 152, 200]
if opt.c3d_model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_name == 'wideresnet':
assert opt.c3d_model_depth in [50]
if opt.c3d_model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, k=opt.wide_resnet_k,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_name == 'resnext':
assert opt.c3d_model_depth in [50, 101, 152]
if opt.c3d_model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut, cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_name == 'preresnet':
assert opt.c3d_model_depth in [18, 34, 50, 101, 152, 200]
if opt.c3d_model_depth == 18:
model = pre_act_resnet.resnet18(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 34:
model = pre_act_resnet.resnet34(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 50:
model = pre_act_resnet.resnet50(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 101:
model = pre_act_resnet.resnet101(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 152:
model = pre_act_resnet.resnet152(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 200:
model = pre_act_resnet.resnet200(num_classes=opt.n_classes, shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_name == 'densenet':
assert opt.c3d_model_depth in [121, 169, 201, 264]
if opt.c3d_model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
elif opt.c3d_model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size, sample_duration=opt.sample_duration,
last_fc=last_fc)
# print(model)
print('loading c3d model from: {}'.format(opt.c3d_model_checkpoint))
model_data = torch.load(opt.c3d_model_checkpoint)
print(model_data['arch'])
assert opt.arch == model_data['arch']
model_state_dict = {}
for k, v in model_data['state_dict'].items():
model_state_dict[k[k.index('.') + 1:]] = v
model.load_state_dict(model_state_dict)
if not opt.no_cuda:
model = model.to(opt.device)
# model = nn.DataParallel(model, device_ids=None)
# print(model)
return model
def get_model(config):
assert config.model in [
'i3d', 'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
print('Initializing {} model (num_classes={})...'.format(
config.model, config.num_classes))
if config.model == 'i3d':
from models.i3d import get_fine_tuning_parameters
model = InceptionI3D(num_classes=config.num_classes,
spatial_squeeze=True,
final_endpoint='logits',
in_channels=3,
dropout_keep_prob=config.dropout_keep_prob)
elif config.model == 'resnet':
assert config.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if config.model_depth == 10:
model = resnet.resnet10(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 18:
model = resnet.resnet18(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 34:
model = resnet.resnet34(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 50:
model = resnet.resnet50(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 101:
model = resnet.resnet101(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 152:
model = resnet.resnet152(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 200:
model = resnet.resnet200(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model == 'wideresnet':
assert config.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if config.model_depth == 50:
model = wide_resnet.resnet50(
num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
k=config.wide_resnet_k,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model == 'resnext':
assert config.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if config.model_depth == 50:
model = resnext.resnet50(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
cardinality=config.resnext_cardinality,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 101:
model = resnext.resnet101(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
cardinality=config.resnext_cardinality,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 152:
model = resnext.resnet152(num_classes=config.num_classes,
shortcut_type=config.resnet_shortcut,
cardinality=config.resnext_cardinality,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model == 'densenet':
assert config.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if config.model_depth == 121:
model = densenet.densenet121(
num_classes=config.num_classes,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 169:
model = densenet.densenet169(
num_classes=config.num_classes,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 201:
model = densenet.densenet201(
num_classes=config.num_classes,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
elif config.model_depth == 264:
model = densenet.densenet264(
num_classes=config.num_classes,
spatial_size=config.spatial_size,
sample_duration=config.sample_duration)
if 'cuda' in config.device:
print('Moving model to CUDA device...')
# Move model to the GPU
model = model.cuda()
if config.model != 'i3d':
model = nn.DataParallel(model, device_ids=None)
if config.checkpoint_path:
print('Loading pretrained model {}'.format(config.checkpoint_path))
assert os.path.isfile(config.checkpoint_path)
checkpoint = torch.load(config.checkpoint_path)
if config.model == 'i3d':
pretrained_weights = checkpoint
else:
pretrained_weights = checkpoint['state_dict']
model.load_state_dict(pretrained_weights)
# Setup finetuning layer for different number of classes
# Note: the DataParallel adds 'module' dict to complicate things...
print('Replacing model logits with {} output classes.'.format(
config.finetune_num_classes))
if config.model == 'i3d':
model.replace_logits(config.finetune_num_classes)
elif config.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
config.finetune_num_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
config.finetune_num_classes)
model.module.fc = model.module.fc.cuda()
# Setup which layers to train
assert config.model in (
'i3d', 'resnet'), 'finetune params not implemented...'
finetune_criterion = config.finetune_prefixes if config.model in (
'i3d', 'resnet') else config.finetune_begin_index
parameters_to_train = get_fine_tuning_parameters(
model, finetune_criterion)
return model, parameters_to_train
else:
raise ValueError('CPU training not supported.')
return model, model.parameters()
def get_network(args):
""" return given network
"""
if args.net == 'vgg16':
from models.vgg import vgg16_bn
net = vgg16_bn()
elif args.net == 'vgg13':
from models.vgg import vgg13_bn
net = vgg13_bn()
elif args.net == 'vgg11':
from models.vgg import vgg11_bn
net = vgg11_bn()
elif args.net == 'vgg19':
from models.vgg import vgg19_bn
net = vgg19_bn()
elif args.net == 'densenet121':
from models.densenet import densenet121
net = densenet121()
elif args.net == 'densenet161':
from models.densenet import densenet161
net = densenet161()
elif args.net == 'densenet169':
from models.densenet import densenet169
net = densenet169()
elif args.net == 'densenet201':
from models.densenet import densenet201
net = densenet201()
elif args.net == 'googlenet':
from models.googlenet import googlenet
net = googlenet()
elif args.net == 'inceptionv3':
from models.inceptionv3 import inceptionv3
net = inceptionv3()
elif args.net == 'inceptionv4':
from models.inceptionv4 import inceptionv4
net = inceptionv4()
elif args.net == 'inceptionresnetv2':
from models.inceptionv4 import inception_resnet_v2
net = inception_resnet_v2()
elif args.net == 'xception':
from models.xception import xception
net = xception()
elif args.net == 'resnet18':
from models.resnet import resnet18
net = resnet18()
elif args.net == 'resnet34':
from models.resnet import resnet34
net = resnet34()
elif args.net == 'resnet50':
from models.resnet import resnet50
net = resnet50()
elif args.net == 'resnet101':
from models.resnet import resnet101
net = resnet101()
elif args.net == 'resnet152':
from models.resnet import resnet152
net = resnet152()
elif args.net == 'preactresnet18':
from models.preactresnet import preactresnet18
net = preactresnet18()
elif args.net == 'preactresnet34':
from models.preactresnet import preactresnet34
net = preactresnet34()
elif args.net == 'preactresnet50':
from models.preactresnet import preactresnet50
net = preactresnet50()
elif args.net == 'preactresnet101':
from models.preactresnet import preactresnet101
net = preactresnet101()
elif args.net == 'preactresnet152':
from models.preactresnet import preactresnet152
net = preactresnet152()
elif args.net == 'resnext50':
from models.resnext import resnext50
net = resnext50()
elif args.net == 'resnext101':
from models.resnext import resnext101
net = resnext101()
elif args.net == 'resnext152':
from models.resnext import resnext152
net = resnext152()
elif args.net == 'shufflenet':
from models.shufflenet import shufflenet
net = shufflenet()
elif args.net == 'shufflenetv2':
from models.shufflenetv2 import shufflenetv2
net = shufflenetv2()
elif args.net == 'squeezenet':
from models.squeezenet import squeezenet
net = squeezenet()
elif args.net == 'mobilenet':
from models.mobilenet import mobilenet
net = mobilenet()
elif args.net == 'mobilenetv2':
from models.mobilenetv2 import mobilenetv2
net = mobilenetv2()
elif args.net == 'nasnet':
from models.nasnet import nasnet
net = nasnet()
elif args.net == 'attention56':
from models.attention import attention56
net = attention56()
elif args.net == 'attention92':
from models.attention import attention92
net = attention92()
elif args.net == 'seresnet18':
from models.senet import seresnet18
net = seresnet18()
elif args.net == 'seresnet34':
from models.senet import seresnet34
net = seresnet34()
elif args.net == 'seresnet50':
from models.senet import seresnet50
net = seresnet50()
elif args.net == 'seresnet101':
from models.senet import seresnet101
net = seresnet101()
elif args.net == 'seresnet152':
from models.senet import seresnet152
net = seresnet152()
elif args.net == 'wideresnet':
from models.wideresidual import wideresnet
net = wideresnet()
elif args.net == 'stochasticdepth18':
from models.stochasticdepth import stochastic_depth_resnet18
net = stochastic_depth_resnet18()
elif args.net == 'stochasticdepth34':
from models.stochasticdepth import stochastic_depth_resnet34
net = stochastic_depth_resnet34()
elif args.net == 'stochasticdepth50':
from models.stochasticdepth import stochastic_depth_resnet50
net = stochastic_depth_resnet50()
elif args.net == 'stochasticdepth101':
from models.stochasticdepth import stochastic_depth_resnet101
net = stochastic_depth_resnet101()
elif args.net == 'normal_resnet':
from models.normal_resnet import resnet18
net = resnet18()
elif args.net == 'hyper_resnet':
from models.hypernet_main import Hypernet_Main
net = Hypernet_Main(
encoder="resnet18",
hypernet_params={'vqvae_dict_size': args.dict_size})
elif args.net == 'normal_resnet_wo_bn':
from models.normal_resnet_wo_bn import resnet18
net = resnet18()
elif args.net == 'hyper_resnet_wo_bn':
from models.hypernet_main import Hypernet_Main
net = Hypernet_Main(
encoder="resnet18_wobn",
hypernet_params={'vqvae_dict_size': args.dict_size})
else:
print('the network name you have entered is not supported yet')
sys.exit()
if args.gpu: #use_gpu
net = net.cuda()
return net
def generate_model(opt):
assert opt.model in [
'resnet', 'resnet_skeleton', 'preresnet', 'wideresnet', 'resnext',
'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnet_skeleton':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet_skeleton import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet_skeleton.resnet_skeleton10(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet_skeleton.resnet_skeleton18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet_skeleton.resnet_skeleton34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet_skeleton.resnet_skeleton50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet_skeleton.resnet_skeleton101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet_skeleton.resnet_skeleton152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet_skeleton.resnet_skeleton200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
if opt.cuda_id is None:
model = model.cuda()
else:
model = model.cuda(opt.cuda_id)
# model = nn.DataParallel(model, device_ids=None)
if opt.cuda_id is None:
model = nn.DataParallel(model, device_ids=None)
else:
model = nn.DataParallel(model, device_ids=[opt.cuda_id])
if opt.pretrain_path:
print(' loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
if opt.model == 'resnet_skeleton':
pretrained_dict = pretrain['state_dict']
model_dict = model.state_dict()
# print('----------------')
# for k, v in pretrained_dict.items():
# if k in model_dict:
# print(k)
# print('----------------')
# pretrained_dict = {k: v for k, v in pretrained_dict.items() if k in model_dict}
pretrained_dict = {
k: v
for k, v in pretrained_dict.items()
if k in model_dict and 'fc' not in k
} ## for concatenate
model_dict.update(pretrained_dict)
model.load_state_dict(model_dict)
else:
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
if opt.cuda_id is None:
model.module.classifier = model.module.classifier.cuda()
else:
model.module.classifier = model.module.classifier.cuda(
opt.cuda_id)
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
if opt.cuda_id is None:
model.module.fc = model.module.fc.cuda()
else:
model.module.fc = model.module.fc.cuda(opt.cuda_id)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
# strip off the 'module.' for each module; this get's added when a model is saved using nn.DataParallel
pretrain['state_dict'] = {
k[7:]: v
for k, v in pretrain['state_dict'].items()
}
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def train_and_val(args):
# logging.basicConfig(filename='example.log', filemode='w', level=logging.DEBUG)
# 第一步，创建一个logger
logger = logging.getLogger('logger')
logger.setLevel(logging.DEBUG) # Log等级总开关
# 第二步，创建一个handler，用于写入日志文件
rq = time.strftime('%Y%m%d%H%M', time.localtime(time.time()))
log_path = os.path.dirname(
os.getcwd()) + '/Diabetic_Reinopathy_Detection/Logs/'
log_name = log_path + rq + '.log'
logfile = log_name
handler = logging.FileHandler(logfile, 'a')
handler.setLevel(logging.DEBUG) # 输出到file的log等级的开关
# 第三步，定义handler的输出格式
formatter = logging.Formatter(
"%(asctime)s - %(filename)s[line:%(lineno)d] - %(levelname)s: %(message)s"
)
handler.setFormatter(formatter)
# 第四步，将logger添加到handler里面
logger.addHandler(handler)
# writer = SummaryWriter(comment='')
train_transform = transforms.Compose([
# transforms.RandomResizedCrop(224), # 从原图像随机切割一张（224， 224）的图像
# transforms.RandomHorizontalFlip(), # 以0.5的概率水平翻转
transforms.RandomVerticalFlip(), # 以0.5的概率垂直翻转
transforms.ColorJitter(0.05, 0.05, 0.05, 0.05), #HSV以及对比度变化
transforms.RandomAffine(45),
transforms.RandomGrayscale(),
transforms.RandomRotation(10), # 在（-10， 10）范围内旋转
transforms.Resize((width, height), interpolation=2),
transforms.ToTensor(),
# transforms.Normalize((.5, .5, .5), (.5, .5, .5))
])
val_transform = transforms.Compose([
transforms.Resize((width, height), interpolation=2),
transforms.ToTensor(),
# transforms.Normalize((.5, .5, .5), (.5, .5, .5))
])
train_ds = DRDataLoader('csv/train_gan.csv',
'/root/lg/dr_datasets_1024/',
transform=train_transform,
train=True)
train_loader = torch.utils.data.DataLoader(train_ds,
batch_size=args.batch_size,
sampler=RandomSampler(train_ds))
val_ds = DRDataLoader('csv/test.csv',
'/root/lg/dr_datasets_1024/',
transform=val_transform,
train=False)
val_loader = torch.utils.data.DataLoader(val_ds,
batch_size=args.test_batch_size)
model = densenet121(pretrained=False)
# fc_features = model.fc.in_features 138448
model.classifier = nn.Linear(model.classifier.in_features, 2, bias=True)
# print(model)
# model = DataParallel(model)
criterion = nn.CrossEntropyLoss(size_average=True)
#
# optimizer = torch.optim.Adam(model.parameters(), lr=args.lr, weight_decay=5e-4)
# Decay LR by a factor of 0.1 every 7 epochs
if not args.disable_cuda and torch.cuda.is_available():
model.cuda()
criterion.cuda()
# Observe that all parameters are being optimized
if args.resume:
if os.path.isfile(os.path.join('./checkpoint', args.resume)):
print("loading checkpoint '{}'".format(args.resume))
checkpoint = torch.load(os.path.join('./checkpoint', args.resume))
args.start_epoch = checkpoint['epoch']
args.acc = checkpoint['acc']
print('epoch', args.start_epoch, 'acc =', args.acc)
model.load_state_dict(checkpoint['state_dict'])
# optimizer.load_state_dict(checkpoint['optimizer'])
# pretrained_dict = checkpoint['state_dict']
# model_dict = model.state_dict()
# pretrained_dict = {k: v for k, v in pretrained_dict.items() if k in model_dict}
# pretrained_dict.pop('classifier.weight')
# pretrained_dict.pop('classifier.bias')
# 2. overwrite entries in the existing state dict
# model_dict.update(pretrained_dict)
# 3. load the new state dict
# model.load_state_dict(model_dict)
logger.debug("loaded checkpoint '{}' (epoch {})".format(
args.resume, checkpoint['epoch']))
else:
print("no checkpoint found at '{}'".format(args.resume))
logger.debug("loaded checkpoint '{}')".format(args.resume))
else:
logger.debug('checkping is none \n')
logger.debug("loaded checkpoint '{}')".format(args.resume))
# for i in enumerate(model.modules()):
# print(i)
# for para in list(model.parameters())[:-3]:
# # print(para)
# para.requires_grad = False
optimizer = torch.optim.SGD(model.parameters(),
lr=args.lr,
momentum=args.momentum,
weight_decay=5e-4)
exp_lr_scheduler = lr_scheduler.StepLR(optimizer, step_size=5, gamma=0.1)
for epoch in range(args.start_epoch + 1, args.epochs):
start = time.time()
model.train()
exp_lr_scheduler.step()
train_loss = []
train_correct = 0
train_total = 0
# Iterate over data.
for idx, (inputs, target, _) in enumerate(train_loader):
if not args.disable_cuda and torch.cuda.is_available():
inputs = inputs.cuda()
target = target.cuda()
inputs = Variable(inputs)
target = Variable(target)
# zero the parameter gradients
optimizer.zero_grad()
outputs = model(inputs)
loss = criterion(outputs, target)
loss.backward()
optimizer.step()
# import numpy as np
# aa = np.log(np.sum(np.exp(outputs.data.cpu().numpy()), axis=1))
# myloss = -outputs.data.cpu().numpy()[0, target.data.cpu().numpy()] + aa
_, preds = torch.max(outputs.data, 1)
# statistics
train_loss.append(loss.data[0])
train_total += inputs.size()[0]
train_correct += (preds == target.data).sum()
# ave_loss = ave_loss * 0.9 + loss.data[0] * 0.1
if idx % args.interval == 0:
print(
'Train Epoch: {} [{}/{} ({:.0f}%)] Loss: {:.5f} Acc: {:.3f}'
.format(epoch, idx * len(inputs),
len(train_loader.dataset),
100. * idx / len(train_loader), loss.data[0],
train_correct * 1.0 / train_total))
s = 'Train Epoch: {} [{}/{} ({:.0f}%)] Loss: {:.5f} Acc: {:.3f}'.format(
epoch, idx * len(inputs), len(train_loader.dataset),
100. * idx / len(train_loader), loss.data[0],
train_correct * 1.0 / train_total)
logger.debug(s)
train_epoch_loss = sum(train_loss) / len(train_loss)
train_epoch_acc = 1.0 * train_correct / train_total
# writer.add_scalar('train' + '/epoch_loss', epoch_loss, epoch)
# writer.add_scalar('train' + '/epoch_acc', epoch_acc, epoch)
print('{} Loss: {:.4f} Acc: {:.4f}'.format('Train', train_epoch_loss,
train_epoch_acc))
logger.debug('{} Loss: {:.4f} Acc: {:.4f}'.format(
'Train', train_epoch_loss, train_epoch_acc))
end = time.time()
training_time = end - start
print('The training time is {:.0f}m {:.0f}s'.format(
training_time // 60, training_time % 60))
model.eval()
val_loss = []
val_correct = 0
val_total = 0
for idx, (inputs, target, _) in enumerate(val_loader):
if not args.disable_cuda and torch.cuda.is_available():
inputs = inputs.cuda()
target = target.cuda()
inputs = Variable(inputs)
target = Variable(target)
outputs = model(inputs)
loss = criterion(outputs, target)
_, preds = torch.max(outputs.data, 1)
# statistics
val_loss.append(loss.data[0])
val_total += inputs.size()[0]
val_correct += (preds == target.data).sum()
print(val_correct, '====', val_total)
#
val_epoch_loss = sum(val_loss) / len(val_loss)
val_epoch_acc = 1.0 * val_correct / val_total
# logger.debug('*********************************************')
print('{} Loss: {:.4f} Acc: {:.4f}'.format('Validate', val_epoch_loss,
val_epoch_acc))
logger.debug('{} Loss: {:.4f} Acc: {:.4f}'.format(
'Validate', val_epoch_loss, val_epoch_acc))
if epoch % 5 == 0 or val_epoch_acc > args.acc:
# if val_epoch_acc > args.acc and train_epoch_acc > 0.5:
logger.debug('Saving...' + str(epoch))
state = {
# 'net': net.module if not args.cuda else net,
'acc': val_epoch_acc,
'epoch': epoch,
'state_dict': model.state_dict(),
'optimizer': optimizer.state_dict()
}
if not os.path.isdir('./checkpoint'):
os.mkdir('./checkpoint')
resume = 'epoch_' + str(epoch) + '_' + args.resume
torch.save(state, os.path.join('./checkpoint', resume))
args.acc = val_epoch_acc
print('Saving...' + str(epoch))
def test(args):
val_transform = transforms.Compose([
transforms.Resize((width, height), interpolation=2),
transforms.ToTensor(),
# transforms.Normalize((.5, .5, .5), (.5, .5, .5))
])
# '/dev/shm/dr_datasets_test/'
val_ds = DRDataLoader('csv/test_gan_350.csv',
'/root/lg/Fake/',
transform=val_transform,
train=False)
val_loader = torch.utils.data.DataLoader(val_ds,
batch_size=args.test_batch_size)
model = densenet121(pretrained=False)
model.classifier = nn.Linear(model.classifier.in_features, 2, bias=True)
criterion = nn.CrossEntropyLoss(size_average=True)
if not args.disable_cuda and torch.cuda.is_available():
model.cuda()
criterion.cuda()
if args.resume:
if os.path.isfile(os.path.join('./checkpoint', args.resume)):
print("loading checkpoint '{}'".format(args.resume))
checkpoint = torch.load(os.path.join('./checkpoint', args.resume))
args.start_epoch = checkpoint['epoch']
args.acc = checkpoint['acc']
print('epoch', args.start_epoch, 'acc =', args.acc)
model.load_state_dict(checkpoint['state_dict'])
else:
print("no checkpoint found at '{}'".format(args.resume))
import csv
start = time.time()
model.eval()
val_loss = []
val_correct = 0
val_total = 0
csvoutPath = './csv/test_gan_350_out_epoch_121.csv'
with open(csvoutPath, 'w', newline='') as f:
f_csv = csv.writer(f)
f_csv.writerow(['file', 'groundtruth', 'predict', 'prob0', 'prob1'])
for idx, (inputs, target, image_name) in enumerate(val_loader):
if not args.disable_cuda and torch.cuda.is_available():
inputs = inputs.cuda()
target = target.cuda()
inputs = Variable(inputs)
target = Variable(target)
outputs = model(inputs)
loss = criterion(outputs, target)
_, preds = torch.max(outputs.data, 1)
# statistics
val_loss.append(loss.data[0])
val_total += inputs.size()[0]
val_correct += (preds == target.data).sum()
h_x = F.softmax(outputs, dim=1).data.squeeze()
# probs, idx = h_x.sort(0, True)
prob, predict = torch.max(outputs.data, 1)
f_csv.writerow([
str(image_name[0]),
target.data.cpu().numpy()[0],
predict.cpu().numpy()[0],
h_x.cpu().numpy()[0],
h_x.cpu().numpy()[1]
])
print(idx)
print(val_correct, '====', val_total)
#
val_epoch_loss = sum(val_loss) / len(val_loss)
val_epoch_acc = 1.0 * val_correct / val_total
# logger.debug('*********************************************')
print('{} Loss: {:.4f} Acc: {:.4f}'.format('Validate', val_epoch_loss,
val_epoch_acc))
# logger.debug('{} Loss: {:.4f} Acc: {:.4f}'.format(
# 'Validate', val_epoch_loss, val_epoch_acc))
end = time.time()
testing_time = end - start
print('The training time is {:.0f}m {:.0f}s'.format(
testing_time // 60, testing_time % 60))
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path, map_location='cpu')
assert opt.arch == pretrain['arch']
from collections import OrderedDict
new_state_dict = OrderedDict()
for k, v in pretrain['state_dict'].items():
name = k[7:] # remove `module.`
new_state_dict[name] = v
# load params
model.load_state_dict(new_state_dict)
#model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def main(df):
# parser = argparse.ArgumentParser()
# parser.add_argument('--batchSz', type=int, default=64)
# parser.add_argument('--nEpochs', type=int, default=300)
# parser.add_argument('--cuda', default=True)
# parser.add_argument('--save')
# parser.add_argument('--seed', type=int, default=1)
# parser.add_argument('--opt', type=str, default='sgd',
# choices=('sgd', 'adam', 'rmsprop'))
# args = parser.parse_args()
# args.cuda = not args.no_cuda and torch.cuda.is_available()
# args.save = args.save or 'work/densenet.base'
setproctitle.setproctitle(args.save)
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
if os.path.exists(args.save):
shutil.rmtree(args.save)
os.makedirs(args.save, exist_ok=True)
normMean = [0.49139968, 0.48215827, 0.44653124]
normStd = [0.24703233, 0.24348505, 0.26158768]
normTransform = transforms.Normalize(normMean, normStd)
trainTransform = transforms.Compose([
transforms.ToPILImage(),
transforms.RandomCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(), normTransform
])
testTransform = transforms.Compose([
transforms.ToPILImage(),
transforms.RandomCrop(224),
transforms.ToTensor(), normTransform
])
df = df.sample(frac=1).reset_index(drop=True)
# df = df[:int(len(df)*0.1)]
train_df = df[:int(len(df) * 0.7)]
test_df = df[int(len(df) * 0.7):].reset_index(drop=True)
kwargs = {'num_workers': 0, 'pin_memory': True} if args.cuda else {}
trainset = CustomDataset(
df=train_df,
root_dir=r'F:\FDAI\NVH\NVH_data\AI_Exterior_Windnoise_image',
transform=trainTransform)
testset = CustomDataset(
df=test_df,
root_dir=r'F:\FDAI\NVH\NVH_data\AI_Exterior_Windnoise_image',
transform=testTransform)
trainLoader = DataLoader(trainset,
batch_size=args.batchSz,
shuffle=True,
**kwargs)
testLoader = DataLoader(testset,
batch_size=args.batchSz,
shuffle=False,
**kwargs)
net = densenet.densenet121()
net = nn.DataParallel(net)
print(' + Number of params: {}'.format(
sum([p.data.nelement() for p in net.parameters()])))
if args.cuda:
net = net.cuda()
if args.opt == 'sgd':
optimizer = optim.SGD(net.parameters(),
lr=1e-1,
momentum=0.9,
weight_decay=1e-4)
elif args.opt == 'adam':
optimizer = optim.Adam(net.parameters(), weight_decay=1e-4)
elif args.opt == 'rmsprop':
optimizer = optim.RMSprop(net.parameters(), weight_decay=1e-4)
trainF = open(os.path.join(args.save, 'train.csv'), 'w')
testF = open(os.path.join(args.save, 'test.csv'), 'w')
for epoch in range(1, args.nEpochs + 1):
adjust_opt(args.opt, optimizer, epoch)
train(args, epoch, net, trainLoader, optimizer, trainF)
test(args, epoch, net, testLoader, optimizer, testF)
torch.save(net, os.path.join(args.save, 'latest.pth'))
os.system('./plot.py {} &'.format(args.save))
trainF.close()
testF.close()
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.my_resnet_v2 import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
pretrain_dict = pretrain['state_dict']
model_dict = model.state_dict()
pretrain_dict = {k: v for k, v in pretrain_dict.items() if k in model_dict}
# print(pretrain_dict.keys())
model_dict.update(pretrain_dict)
model.load_state_dict(model_dict)
# model.load_state_dict(pretrain['state_dict']
# if opt.model == 'densenet':
# model.module.classifier = nn.Linear(
# model.module.classifier.in_features, opt.n_finetune_classes)
# model.module.classifier = model.module.classifier.cuda()
# else:
# model.module.fc = nn.Linear(model.module.fc.in_features,
# opt.n_finetune_classes)
# model.module.fc = model.module.fc.cuda()
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
pretrain_dict = pretrain['state_dict']
model_dict = model.state_dict()
pretrain_dict = {k: v for k, v in pretrain_dict.items() if k in model_dict}
model_dict.update(pretrain_dict)
model.load_state_dict(model_dict)
# model.load_state_dict(pretrain['state_dict']
# if opt.model == 'densenet':
# model.classifier = nn.Linear(
# model.classifier.in_features, opt.n_finetune_classes)
# else:
# model.fc = nn.Linear(model.fc.in_features,
# opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()#如果没有pretrain_path就输出模型的所有参数
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
###################################################################
# ResNet
###################################################################
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
###################################################################
# Wider ResNet
###################################################################
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
###################################################################
# ResNext
###################################################################
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
###################################################################
# Pre-ResNet
###################################################################
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
###################################################################
# DenseNet
###################################################################
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
###################################################################
# Finalizing the model
###################################################################
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=opt.device_ids)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch'] # ensure that pretrain model is the same architecture
model.load_state_dict(pretrain['state_dict'])
# change the fc layer output size
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features, opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
#
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(
model.classifier.in_features, opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'resnext_fa', 'densenet', 'p3d'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'p3d':
assert opt.model_depth in [50, 101, 152]
if opt.model_depth == 50:
model = p3d.P3D63(num_classes=opt.n_classes)
elif opt.model_depth == 101:
model = p3d.P3D131(num_classes=opt.n_classes)
elif opt.model_depth == 152:
model = p3d.P3D199(num_classes=opt.n_classes)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext_fa import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext_fa':
assert opt.model_depth in [50, 101, 152]
from models.resnext_fa import get_fine_tuning_parameters, get_fine_tuning_parameters_fa
if opt.model_depth == 50:
model = resnext_fa.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext_fa.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext_fa.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
#pdb.set_trace();
#assert opt.arch == pretrain['arch']
model_dict = model.state_dict();
#pdb.set_trace();
model_dict.update(pretrain['state_dict']);
model.load_state_dict(model_dict);
#model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features, opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
# do not need to add new fc layer when finetuning model has the same class num
elif (opt.n_classes != opt.n_finetune_classes):
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
if (opt.model == 'resnext_fa'):
parameters = get_fine_tuning_parameters_fa(model, opt.learning_rate)
else:
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(
model.classifier.in_features, opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet',
'mobilenet', 'mobilenetv2'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'mobilenet':
from models.mobilenet import get_fine_tuning_parameters
model = mobilenet.get_model(num_classes=opt.n_classes,
sample_size=opt.sample_size,
width_mult=opt.width_mult)
elif opt.model == 'mobilenetv2':
from models.mobilenetv2 import get_fine_tuning_parameters
model = mobilenetv2.get_model(num_classes=opt.n_classes,
sample_size=opt.sample_size,
width_mult=opt.width_mult)
if not opt.no_cuda:
if not opt.no_cuda_predict:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
print("Pretrain arch", pretrain['arch'])
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
ft_begin_index = opt.ft_begin_index
if opt.model in [
'mobilenet', 'mobilenetv2', 'shufflenet', 'shufflenetv2'
]:
model.module.classifier = nn.Sequential(
nn.Dropout(0.9),
nn.Linear(model.module.classifier[1].in_features,
opt.n_finetune_classes))
model.module.classifier = model.module.classifier.cuda()
ft_begin_index = 'complete' if ft_begin_index == 0 else 'last_layer'
elif opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
print("Finetuning at:", ft_begin_index)
parameters = get_fine_tuning_parameters(model, ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
ft_begin_index = opt.ft_begin_index
if opt.model in [
'mobilenet', 'mobilenetv2', 'shufflenet', 'shufflenetv2'
]:
model.module.classifier = nn.Sequential(
nn.Dropout(0.9),
nn.Linear(model.module.classifier[1].in_features,
opt.n_finetune_classes))
model.module.classifier = model.module.classifier.cuda()
ft_begin_index = 'complete' if ft_begin_index == 0 else 'last_layer'
elif opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
print("Finetuning at:", ft_begin_index)
parameters = get_fine_tuning_parameters(model, ft_begin_index)
return model, parameters
return model, model.parameters()
def get_network(args, use_gpu=True):
""" return given network
"""
if args.net == 'vgg16':
from models.vgg import vgg16_bn
net = vgg16_bn()
elif args.net == 'vgg13':
from models.vgg import vgg13_bn
net = vgg13_bn()
elif args.net == 'vgg11':
from models.vgg import vgg11_bn
net = vgg11_bn()
elif args.net == 'vgg19':
from models.vgg import vgg19_bn
net = vgg19_bn()
elif args.net == 'densenet121':
from models.densenet import densenet121
net = densenet121()
elif args.net == 'densenet161':
from models.densenet import densenet161
net = densenet161()
elif args.net == 'densenet169':
from models.densenet import densenet169
net = densenet169()
elif args.net == 'densenet201':
from models.densenet import densenet201
net = densenet201()
elif args.net == 'googlenet':
from models.googlenet import googlenet
net = googlenet()
elif args.net == 'inceptionv3':
from models.inceptionv3 import inceptionv3
net = inceptionv3()
elif args.net == 'inceptionv4':
from models.inceptionv4 import inceptionv4
net = inceptionv4()
elif args.net == 'inceptionresnetv2':
from models.inceptionv4 import inception_resnet_v2
net = inception_resnet_v2()
elif args.net == 'xception':
from models.xception import xception
net = xception()
elif args.net == 'resnet18':
from models.resnet import resnet18
net = resnet18()
elif args.net == 'resnet34':
from models.resnet import resnet34
net = resnet34()
elif args.net == 'resnet50':
from models.resnet import resnet50
net = resnet50()
elif args.net == 'resnet101':
from models.resnet import resnet101
net = resnet101()
elif args.net == 'resnet152':
from models.resnet import resnet152
net = resnet152()
elif args.net == 'preactresnet18':
from models.preactresnet import preactresnet18
net = preactresnet18()
elif args.net == 'preactresnet34':
from models.preactresnet import preactresnet34
net = preactresnet34()
elif args.net == 'preactresnet50':
from models.preactresnet import preactresnet50
net = preactresnet50()
elif args.net == 'preactresnet101':
from models.preactresnet import preactresnet101
net = preactresnet101()
elif args.net == 'preactresnet152':
from models.preactresnet import preactresnet152
net = preactresnet152()
elif args.net == 'resnext50':
from models.resnext import resnext50
net = resnext50()
elif args.net == 'resnext101':
from models.resnext import resnext101
net = resnext101()
elif args.net == 'resnext152':
from models.resnext import resnext152
net = resnext152()
elif args.net == 'shufflenet':
from models.shufflenet import shufflenet
net = shufflenet()
elif args.net == 'shufflenetv2':
from models.shufflenetv2 import shufflenetv2
net = shufflenetv2()
elif args.net == 'squeezenet':
from models.squeezenet import squeezenet
net = squeezenet()
elif args.net == 'mobilenet':
from models.mobilenet import mobilenet
net = mobilenet()
elif args.net == 'mobilenetv2':
from models.mobilenetv2 import mobilenetv2
net = mobilenetv2()
elif args.net == 'nasnet':
from models.nasnet import nasnet
net = nasnet()
elif args.net == 'attention56':
from models.attention import attention56
net = attention56()
elif args.net == 'attention92':
from models.attention import attention92
net = attention92()
elif args.net == 'seresnet18':
from models.senet import seresnet18
net = seresnet18()
elif args.net == 'seresnet34':
from models.senet import seresnet34
net = seresnet34()
elif args.net == 'seresnet50':
from models.senet import seresnet50
net = seresnet50()
elif args.net == 'seresnet101':
from models.senet import seresnet101
net = seresnet101()
elif args.net == 'seresnet152':
from models.senet import seresnet152
net = seresnet152()
else:
print('the network name you have entered is not supported yet')
sys.exit()
if use_gpu:
net = net.cuda()
return net
def get_network(netname, use_gpu=True):
""" return given network
"""
if netname == 'vgg16':
from models.vgg import vgg16_bn #!
net = vgg16_bn()
elif netname == 'vgg16_cbn':
from models.vgg_nobn import vgg16_cbn #!
net = vgg16_cbn()
elif netname == 'vgg11':
from models.vgg import vgg11_bn #!
net = vgg11_bn()
elif netname == 'vgg11_cbn':
from models.vgg_nobn import vgg11_cbn #!
net = vgg11_cbn()
elif netname == 'vgg11_nobn':
from models.vgg_nobn import vgg11_nobn #!
net = vgg11_nobn()
elif netname == 'vgg16_nobn':
from models.vgg_nobn import vgg16_nobn #!
net = vgg16_nobn()
elif netname == 'resnet18':
from models.resnet import resnet18
net = resnet18()
elif netname == 'resnet50':
from models.resnet import resnet50
net = resnet50()
elif netname == 'resnet101':
from models.resnet import resnet101
net = resnet101()
elif netname == 'resnet18_nobn':
from models.resnet_nobn import resnet18_nobn
net = resnet18_nobn()
elif netname == 'resnet18_fixup':
from models.resnet_fixup import resnet18
net = resnet18()
elif netname == 'resnet50_fixup':
from models.resnet_fixup import resnet50
net = resnet50()
elif netname == 'resnet18_cbn':
from models.resnet_nobn import resnet18_cbn
net = resnet18_cbn()
elif netname == 'resnet50_cbn':
from models.resnet_nobn import resnet50_cbn
net = resnet50_cbn()
elif netname == 'resnet50_nobn':
from models.resnet_nobn import resnet50_nobn
net = resnet50_nobn()
elif netname == 'resnet101_cbn':
from models.resnet_nobn import resnet101_cbn
net = resnet101_cbn()
elif netname == 'densenet121':
from models.densenet import densenet121
net = densenet121()
elif netname == 'densenet121_cbn':
from models.densenet_nobn import densenet121
net = densenet121()
elif netname == 'shufflenetv2':
from models.shufflenetv2 import shufflenetv2
net = shufflenetv2()
elif netname == 'shufflenetv2_cbn':
from models.shufflenetv2_nobn import shufflenetv2_cbn
net = shufflenetv2_cbn()
elif netname == 'shufflenetv2_nobn':
from models.shufflenetv2_nobn import shufflenetv2_nobn
net = shufflenetv2_nobn()
elif netname == 'squeezenet':
from models.squeezenet import squeezenet
net = squeezenet()
elif netname == 'squeezenet_nobn':
from models.squeezenet_nobn import squeezenet_nobn
net = squeezenet_nobn()
elif netname == 'squeezenet_cbn':
from models.squeezenet_nobn import squeezenet_cbn
net = squeezenet_cbn()
elif netname == 'seresnet18':
from models.senet import seresnet18
net = seresnet18()
elif netname == 'seresnet50':
from models.senet import seresnet50
net = seresnet50()
elif netname == 'seresnet18_cbn':
from models.senet_nobn import seresnet18
net = seresnet18()
elif netname == 'seresnet50_cbn':
from models.senet_nobn import seresnet50
net = seresnet50()
elif netname == 'fixup_cbn':
from models.fixup_resnet_cifar import fixup_resnet56
net = fixup_resnet56(cbn=True)
elif netname == 'fixup':
from models.fixup_resnet_cifar import fixup_resnet56
net = fixup_resnet56()
elif netname == 'mobilenetv2':
from models.mobilenetv2 import mobilenetv2
net = mobilenetv2()
elif netname == 'mobilenetv2_cbn':
from models.mobilenetv2_nobn import mobilenetv2
net = mobilenetv2()
else:
print(netname)
print('the network name you have entered is not supported yet')
sys.exit()
if use_gpu:
# net = torch.nn.parallel.DataParallel(net)
net = net.cuda()
return net
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet', 'i3d',
'i3dv2'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == "i3d":
from models.i3dpt import get_fine_tuning_parameters
model = i3dpt.I3D(num_classes=opt.n_classes, dropout_prob=0.5)
elif opt.model == "i3dv2":
from models.I3D_Pytorch import get_fine_tuning_parameters
model = I3D_Pytorch.I3D(num_classes=opt.n_classes,
dropout_keep_prob=0.5)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
if opt.model != "i3d" and opt.model != "i3dv2":
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
else:
pretrain = {"module." + k: v for k, v in pretrain.items()}
model_dict = model.state_dict()
model_dict.update(pretrain)
model.load_state_dict(model_dict)
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def get_network(args, use_gpu=True):
""" return given network
"""
if args.net == 'mobilenet':
from models.mobilenet import mobilenet
net = mobilenet(args)
elif args.net == 'mobilenetv2':
from models.mobilenetv2 import mobilenetv2
net = mobilenetv2(args)
elif args.net == 'vgg13':
from models.vgg import vgg13_bn
net = vgg13_bn(args)
elif args.net == 'vgg11':
from models.vgg import vgg11_bn
net = vgg11_bn(args)
elif args.net == 'vgg19':
# from models.vgg import vgg19_bn
# net = vgg19_bn(args)
from torchvision.models import vgg19_bn
import torch.nn as nn
net = vgg19_bn(pretrained=True)
net.classifier[6] = nn.Linear(4096, args.nc)
elif args.net == 'densenet121':
from models.densenet import densenet121
net = densenet121(args)
elif args.net == 'densenet161':
from models.densenet import densenet161
net = densenet161(args)
elif args.net == 'densenet169':
from models.densenet import densenet169
net = densenet169(args)
elif args.net == 'densenet201':
from models.densenet import densenet201
net = densenet201(args)
elif args.net == 'googlenet':
from models.googlenet import googlenet
net = googlenet(args)
elif args.net == 'inceptionv3':
from models.inceptionv3 import inceptionv3
net = inceptionv3(args)
elif args.net == 'inceptionv4':
from models.inceptionv4 import inceptionv4
net = inceptionv4(args)
elif args.net == 'inceptionresnetv2':
from models.inceptionv4 import inception_resnet_v2
net = inception_resnet_v2(args)
elif args.net == 'xception':
from models.xception import xception
net = xception(args)
elif args.net == 'resnet18':
# from models.resnet import resnet18
# net = resnet18(args)
from torchvision.models import resnet18
import torch.nn as nn
net = resnet18(pretrained=True)
net.fc = nn.Linear(512, args.nc)
elif args.net == 'resnet34':
from models.resnet import resnet34
net = resnet34(args)
elif args.net == 'resnet50':
from models.resnet import resnet50
net = resnet50(args)
elif args.net == 'resnet101':
from models.resnet import resnet101
net = resnet101(args)
elif args.net == 'resnet152':
from models.resnet import resnet152
net = resnet152(args)
elif args.net == 'preactresnet18':
from models.preactresnet import preactresnet18
net = preactresnet18(args)
elif args.net == 'preactresnet34':
from models.preactresnet import preactresnet34
net = preactresnet34(args)
elif args.net == 'preactresnet50':
from models.preactresnet import preactresnet50
net = preactresnet50(args)
elif args.net == 'preactresnet101':
from models.preactresnet import preactresnet101
net = preactresnet101(args)
elif args.net == 'preactresnet152':
from models.preactresnet import preactresnet152
net = preactresnet152(args)
elif args.net == 'resnext50':
from models.resnext import resnext50
net = resnext50(args)
elif args.net == 'resnext101':
from models.resnext import resnext101
net = resnext101(args)
elif args.net == 'resnext152':
from models.resnext import resnext152
net = resnext152(args)
elif args.net == 'shufflenet':
from models.shufflenet import shufflenet
net = shufflenet(args)
elif args.net == 'shufflenetv2':
from models.shufflenetv2 import shufflenetv2
net = shufflenetv2(args)
elif args.net == 'squeezenet':
from models.squeezenet import squeezenet
net = squeezenet(args)
elif args.net == 'mobilenet':
from models.mobilenet import mobilenet
net = mobilenet(args)
elif args.net == 'mobilenetv2':
from models.mobilenetv2 import mobilenetv2
net = mobilenetv2(args)
elif args.net == 'mobilenetv3':
from models.mobilenetv3 import mobileNetv3
net = mobileNetv3(args)
elif args.net == 'mobilenetv3_l':
from models.mobilenetv3 import mobileNetv3
net = mobileNetv3(args, mode='large')
elif args.net == 'mobilenetv3_s':
from models.mobilenetv3 import mobileNetv3
net = mobileNetv3(args, mode='small')
elif args.net == 'nasnet':
from models.nasnet import nasnetalarge
net = nasnetalarge(args)
elif args.net == 'attention56':
from models.attention import attention56
net = attention56(args)
elif args.net == 'attention92':
from models.attention import attention92
net = attention92(args)
elif args.net == 'seresnet18':
from models.senet import seresnet18
net = seresnet18(args)
elif args.net == 'seresnet34':
from models.senet import seresnet34
net = seresnet34(args)
elif args.net == 'seresnet50':
from models.senet import seresnet50
net = seresnet50(args)
elif args.net == 'seresnet101':
from models.senet import seresnet101
net = seresnet101(args)
elif args.net == 'seresnet152':
from models.senet import seresnet152
net = seresnet152(args)
elif args.net.lower() == 'sqnxt_23_1x':
from models.SqueezeNext import SqNxt_23_1x
net = SqNxt_23_1x(args)
elif args.net.lower() == 'sqnxt_23_1xv5':
from models.SqueezeNext import SqNxt_23_1x_v5
net = SqNxt_23_1x_v5(args)
elif args.net.lower() == 'sqnxt_23_2x':
from models.SqueezeNext import SqNxt_23_2x
net = SqNxt_23_2x(args)
elif args.net.lower() == 'sqnxt_23_2xv5':
from models.SqueezeNext import SqNxt_23_2x_v5
net = SqNxt_23_2x_v5(args)
elif args.net.lower() == 'mnasnet':
# from models.MnasNet import mnasnet
# net = mnasnet(args)
from models.nasnet_mobile import nasnet_Mobile
net = nasnet_Mobile(args)
elif args.net == 'efficientnet_b0':
from models.efficientnet import efficientnet_b0
net = efficientnet_b0(args)
elif args.net == 'efficientnet_b1':
from models.efficientnet import efficientnet_b1
net = efficientnet_b1(args)
elif args.net == 'efficientnet_b2':
from models.efficientnet import efficientnet_b2
net = efficientnet_b2(args)
elif args.net == 'efficientnet_b3':
from models.efficientnet import efficientnet_b3
net = efficientnet_b3(args)
elif args.net == 'efficientnet_b4':
from models.efficientnet import efficientnet_b4
net = efficientnet_b4(args)
elif args.net == 'efficientnet_b5':
from models.efficientnet import efficientnet_b5
net = efficientnet_b5(args)
elif args.net == 'efficientnet_b6':
from models.efficientnet import efficientnet_b6
net = efficientnet_b6(args)
elif args.net == 'efficientnet_b7':
from models.efficientnet import efficientnet_b7
net = efficientnet_b7(args)
elif args.net == 'mlp':
from models.mlp import MLPClassifier
net = MLPClassifier(args)
elif args.net == 'alexnet':
from torchvision.models import alexnet
import torch.nn as nn
net = alexnet(pretrained=True)
net.classifier[6] = nn.Linear(4096, args.nc)
elif args.net == 'lambda18':
from models._lambda import LambdaResnet18
net = LambdaResnet18(num_classes=args.nc, channels=args.cs)
elif args.net == 'lambda34':
from models._lambda import LambdaResnet34
net = LambdaResnet34(num_classes=args.nc, channels=args.cs)
elif args.net == 'lambda50':
from models._lambda import LambdaResnet50
net = LambdaResnet50(num_classes=args.nc, channels=args.cs)
elif args.net == 'lambda101':
from models._lambda import LambdaResnet101
net = LambdaResnet101(num_classes=args.nc)
elif args.net == 'lambda152':
from models._lambda import LambdaResnet152
net = LambdaResnet152(num_classes=args.nc, channels=args.cs)
else:
print('the network name you have entered is not supported yet')
sys.exit()
if use_gpu:
net = net.cuda()
return net
def get_model(class_num):
if (MODEL_TYPE == 'alexnet'):
model = alexnet.alexnet(pretrained=FINETUNE)
elif (MODEL_TYPE == 'vgg'):
if (MODEL_DEPTH_OR_VERSION == 11):
model = vgg.vgg11(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 13):
model = vgg.vgg13(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 16):
model = vgg.vgg16(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 19):
model = vgg.vgg19(pretrained=FINETUNE)
else:
print('Error : VGG should have depth of either [11, 13, 16, 19]')
sys.exit(1)
elif (MODEL_TYPE == 'squeezenet'):
if (MODEL_DEPTH_OR_VERSION == 0 or MODEL_DEPTH_OR_VERSION == 'v0'):
model = squeezenet.squeezenet1_0(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 1 or MODEL_DEPTH_OR_VERSION == 'v1'):
model = squeezenet.squeezenet1_1(pretrained=FINETUNE)
else:
print('Error : Squeezenet should have version of either [0, 1]')
sys.exit(1)
elif (MODEL_TYPE == 'resnet'):
if (MODEL_DEPTH_OR_VERSION == 18):
model = resnet.resnet18(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 34):
model = resnet.resnet34(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 50):
model = resnet.resnet50(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 101):
model = resnet.resnet101(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 152):
model = resnet.resnet152(pretrained=FINETUNE)
else:
print(
'Error : Resnet should have depth of either [18, 34, 50, 101, 152]'
)
sys.exit(1)
elif (MODEL_TYPE == 'densenet'):
if (MODEL_DEPTH_OR_VERSION == 121):
model = densenet.densenet121(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 169):
model = densenet.densenet169(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 161):
model = densenet.densenet161(pretrained=FINETUNE)
elif (MODEL_DEPTH_OR_VERSION == 201):
model = densenet.densenet201(pretrained=FINETUNE)
else:
print(
'Error : Densenet should have depth of either [121, 169, 161, 201]'
)
sys.exit(1)
elif (MODEL_TYPE == 'inception'):
if (MODEL_DEPTH_OR_VERSION == 3 or MODEL_DEPTH_OR_VERSION == 'v3'):
model = inception.inception_v3(pretrained=FINETUNE)
else:
print('Error : Inception should have version of either [3, ]')
sys.exit(1)
else:
print(
'Error : Network should be either [alexnet / squeezenet / vgg / resnet / densenet / inception]'
)
sys.exit(1)
if (MODEL_TYPE == 'alexnet' or MODEL_TYPE == 'vgg'):
num_ftrs = model.classifier[6].in_features
feature_model = list(model.classifier.children())
feature_model.pop()
feature_model.append(nn.Linear(num_ftrs, class_num))
model.classifier = nn.Sequential(*feature_model)
elif (MODEL_TYPE == 'resnet' or MODEL_TYPE == 'inception'):
num_ftrs = model.fc.in_features
model.fc = nn.Linear(num_ftrs, class_num)
elif (MODEL_TYPE == 'densenet'):
num_ftrs = model.classifier.in_features
model.classifier = nn.Linear(num_ftrs, class_num)
return model
import copy
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# alexnet = models.alexnet(pretrained=True)
# resnet18 = models.resnet18(pretrained=True).to(device)
# resnet18.device = device
# vgg16 = models.vgg16(pretrained=True)
# densenet = models.densenet161(pretrained=True)
# squeezenet = models.squeezenet1_0(pretrained=True)
resnet18 = resnet.resnet18(pretrained=True).to(device)
resnet18.device = device
resnet18.name = "resnet18"
densenet = densenet.densenet121(pretrained=True).to(device)
densenet.device = device
densenet.name = "densenet"
vgg16 = vgg.vgg16_bn(pretrained=True).to(device)
vgg16.device = device
vgg16.name = "vgg16"
net_list = [resnet18, densenet, vgg16]
normalize = transforms.Normalize(mean=[0.4914, 0.4822, 0.4465],
std=[0.2023, 0.1994, 0.2010])
transform = transforms.Compose([transforms.ToTensor(), normalize])
testset = datasets.CIFAR10(root="./data",
train=False,
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
model_type=opt.model_type)
elif opt.model_depth == 101:
model = resnet.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
model_type=opt.model_type)
elif opt.model_depth == 152:
model = resnet.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
model_type=opt.model_type)
elif opt.model_depth == 200:
model = resnet.resnet200(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration,
model_type=opt.model_type)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
import os
# os.environ['CUDA_VISIBLE_DEVICES'] = f'{opt.cuda_id}'
model = model.cuda(device=opt.cuda_id)
model = nn.DataParallel(model, device_ids=[0]) # CUDA change
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
print(pretrain['arch'])
arch = f'{opt.model}-{opt.model_depth}'
# arch = opt.model + '-' + opt.model_depth
print(arch)
assert arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features,
opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda(
device=opt.cuda_id)
# elif opt.use_quadriplet:
# model = EmbeddingModel(model, opt.n_finetune_classes, not opt.no_cuda, opt.cuda_id)
else:
model.module.fc = nn.Sequential(
nn.Dropout(0.4),
nn.Linear(model.module.fc.in_features, 512), nn.ReLU6(),
nn.Dropout(0.4), nn.Linear(512, 128), nn.ReLU6(),
nn.Linear(128,
opt.n_finetune_classes)).cuda(device=opt.cuda_id)
# model.module.fc = nn.Linear(model.module.fc.in_features,
# opt.n_finetune_classes)
# model.module.fc = model.module.fc.cuda(device=opt.cuda_id)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
print(len(list(parameters)), 'params to fine tune')
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(model.classifier.in_features,
opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
def generate_model(opt):
assert opt.model in [
'resnet', 'preresnet', 'wideresnet', 'resnext', 'densenet'
]
if opt.model == 'resnet':
assert opt.model_depth in [10, 18, 34, 50, 101, 152, 200]
from models.resnet import get_fine_tuning_parameters
if opt.model_depth == 10:
model = resnet.resnet10(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 18:
model = resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'wideresnet':
assert opt.model_depth in [50]
from models.wide_resnet import get_fine_tuning_parameters
if opt.model_depth == 50:
model = wide_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
k=opt.wide_resnet_k,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'resnext':
assert opt.model_depth in [50, 101, 152]
from models.resnext import get_fine_tuning_parameters
if opt.model_depth == 50:
model = resnext.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = resnext.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = resnext.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
cardinality=opt.resnext_cardinality,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'preresnet':
assert opt.model_depth in [18, 34, 50, 101, 152, 200]
from models.pre_act_resnet import get_fine_tuning_parameters
if opt.model_depth == 18:
model = pre_act_resnet.resnet18(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 34:
model = pre_act_resnet.resnet34(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 50:
model = pre_act_resnet.resnet50(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 101:
model = pre_act_resnet.resnet101(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 152:
model = pre_act_resnet.resnet152(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 200:
model = pre_act_resnet.resnet200(
num_classes=opt.n_classes,
shortcut_type=opt.resnet_shortcut,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model == 'densenet':
assert opt.model_depth in [121, 169, 201, 264]
from models.densenet import get_fine_tuning_parameters
if opt.model_depth == 121:
model = densenet.densenet121(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 169:
model = densenet.densenet169(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 201:
model = densenet.densenet201(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
elif opt.model_depth == 264:
model = densenet.densenet264(
num_classes=opt.n_classes,
sample_size=opt.sample_size,
sample_duration=opt.sample_duration)
if not opt.no_cuda:
model = model.cuda()
model = nn.DataParallel(model, device_ids=None)
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.module.classifier = nn.Linear(
model.module.classifier.in_features, opt.n_finetune_classes)
model.module.classifier = model.module.classifier.cuda()
else:
model.module.fc = nn.Linear(model.module.fc.in_features,
opt.n_finetune_classes)
model.module.fc = model.module.fc.cuda()
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
else:
if opt.pretrain_path:
print('loading pretrained model {}'.format(opt.pretrain_path))
pretrain = torch.load(opt.pretrain_path)
assert opt.arch == pretrain['arch']
model.load_state_dict(pretrain['state_dict'])
if opt.model == 'densenet':
model.classifier = nn.Linear(
model.classifier.in_features, opt.n_finetune_classes)
else:
model.fc = nn.Linear(model.fc.in_features,
opt.n_finetune_classes)
parameters = get_fine_tuning_parameters(model, opt.ft_begin_index)
return model, parameters
return model, model.parameters()
