-
Notifications
You must be signed in to change notification settings - Fork 6
/
utils.py
150 lines (118 loc) · 5.26 KB
/
utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
import torch
import torch.nn as nn
import torch.nn.functional as F
from torchvision import transforms
from torchvision.datasets import ImageFolder
from network import models
def get_data_info(args):
if args.target in ['amazon', 'dslr', 'webcam']: # Office-31
resnet_type = 50
num_classes = 31
else: # Office-Home
resnet_type = 50
num_classes = 65
return num_classes, resnet_type
def get_dataset(domain_name, db_path):
if domain_name in ['amazon', 'dslr', 'webcam']: # OFFICE-31
data_transforms = {
'train': transforms.Compose([
transforms.Resize([256, 256]),
transforms.RandomCrop(224),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
'test': transforms.Compose([
transforms.Resize([224, 224]),
transforms.CenterCrop((224, 224)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
}
tr_dataset = ImageFolder(db_path + '/office31/' + domain_name + '/', data_transforms['train'])
te_dataset = ImageFolder(db_path + '/office31/' + domain_name + '/', data_transforms['test'])
elif domain_name in ['art', 'product', 'clipart', 'realworld']: # Office-Home
data_transforms = {
'train': transforms.Compose([
transforms.Resize([256, 256]),
transforms.RandomCrop((224, 224)),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
'test': transforms.Compose([
transforms.Resize([256, 256]),
transforms.CenterCrop((224, 224)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
]),
}
tr_dataset = ImageFolder(db_path + '/OfficeHome/OfficeHomeDataset/' + domain_name + '/', data_transforms['train'])
te_dataset = ImageFolder(db_path + '/OfficeHome/OfficeHomeDataset/' + domain_name + '/', data_transforms['test'])
else:
raise ValueError('Domain %s is not found!' % domain_name)
print('{} train set size: {}'.format(domain_name, len(tr_dataset)))
print('{} test set size: {}'.format(domain_name, len(te_dataset)))
return tr_dataset, te_dataset
def get_train_info():
lr = 0.002
l2_decay = 5e-4
momentum = 0.9
print('lr, l2_decay, momentum:', lr, l2_decay, momentum)
return lr, l2_decay, momentum
def get_net_info(num_classes):
net = nn.DataParallel(models.ResNet50().encoder).cuda()
head = nn.DataParallel(models.Head()).cuda()
classifier = nn.DataParallel(nn.Linear(256, num_classes)).cuda()
emp_learner = nn.DataParallel(models.EmpLearner()).cuda()
return net, head, classifier, emp_learner
def load_net(args, net, head, classifier, root):
print("Load pre-trained model !")
save_folder = root + args.baseline_path
net.module.load_state_dict(torch.load(save_folder + '/net.pt'), strict=False)
head.module.load_state_dict(torch.load(save_folder + '/head.pt'), strict=False)
classifier.module.load_state_dict(torch.load(save_folder + '/classifier.pt'), strict=False)
return net, head, classifier
def set_model_mode(mode, models):
for model in models:
if mode == 'train':
model.train()
else:
model.eval()
def sample_wise_loss(pred, y_a, y_b, lam):
pred = F.log_softmax(pred.unsqueeze(dim=0), dim=-1)
return lam * F.nll_loss(pred, y_a.unsqueeze(dim=0)).cuda() + (1 - lam) * F.nll_loss(pred, y_b.unsqueeze(dim=0)).cuda()
def get_top2(q):
topk_prob, topk_label = torch.topk(F.softmax(q, dim=1), k=2)
topk_label = topk_label.squeeze()
top1_label, top2_label = topk_label.t()[0], topk_label.t()[1]
top1_prob = topk_prob.squeeze().t()[0]
return top1_label.detach(), top2_label.detach(), top1_prob.detach()
def get_vicinal_instance(src_imgs, tgt_imgs, emp, num_of_samples):
mixed_input = []
for i in range(num_of_samples):
mixed_input.append(src_imgs[i] * emp[i] + tgt_imgs[i] * (1 - emp[i]))
mixed_input = torch.stack(mixed_input)
return mixed_input
class EntropyLoss(nn.Module):
def __init__(self):
super(EntropyLoss, self).__init__()
def forward(self, x):
out = F.softmax(x, dim=1)
loss = torch.mean(torch.sum(out * (torch.log(out + 1e-5)), 1))
return loss
def evaluate(model, loader):
total, correct = 0, 0
set_model_mode('eval', model)
with torch.no_grad():
for step, tgt_data in enumerate(loader):
tgt_imgs, tgt_labels = tgt_data
tgt_imgs, tgt_labels = tgt_imgs.cuda(non_blocking=True), tgt_labels.cuda(non_blocking=True)
tgt_pred = model(tgt_imgs)
pred = tgt_pred.argmax(dim=1, keepdim=True)
correct += pred.eq(tgt_labels.long().view_as(pred)).sum().item()
total += tgt_labels.size(0)
acc = (correct / total) * 100
print('Accuracy: {:.2f}%'.format(acc))
set_model_mode('train', model)
return acc