-
Notifications
You must be signed in to change notification settings - Fork 0
/
FSU.py
197 lines (168 loc) · 7.72 KB
/
FSU.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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
import torch
import torch.nn as nn
import torch.optim as optim
import torch.backends.cudnn as cudnn
from torch.optim.lr_scheduler import MultiStepLR
from torchvision import datasets, transforms
from scipy.spatial import distance
from scipy.stats import chi2
from scipy import stats
from scipy.optimize import curve_fit
from sklearn.preprocessing import normalize
import argparse
import os
import csv
import math
import pandas as pd
import numpy as np
import resource
from collections import OrderedDict
from model import resnet
from model import resnet18
from model import densenet_BC
from model import vgg
from model import mobilenet
from model import efficientnet
from model import wrn
from model import convmixer
from utils import data as dataset
from utils import crl_utils
from utils import metrics
from utils import utils
import train_cvpr
from utils.data_utils import *
rlimit = resource.getrlimit(resource.RLIMIT_NOFILE)
resource.setrlimit(resource.RLIMIT_NOFILE, (2048, rlimit[1]))
def csv_writter(path, dic, start):
if os.path.isdir(path) == False: os.makedirs(path)
os.chdir(path)
if start == 1:
mode = 'w'
else:
mode = 'a'
with open('logs.csv', mode) as csvfile:
writer = csv.writer(csvfile, delimiter=",")
if start == 1:
writer.writerow(dic.keys())
writer.writerow([elem["string"] for elem in dic.values()])
class Counter(dict):
def __missing__(self, key):
return None
def dist_feature(mode, features, target_labels): ##function to compute FSU
features_locs = []
for lab in np.unique(target_labels):
features_locs.append(np.where(target_labels == lab)[0])
if mode == 'intra':
if isinstance(features, torch.Tensor):
intrafeatures = features.detach().cpu().numpy()
else:
intrafeatures = features
intra_dists = []
for loc in features_locs:
c_dists = distance.cdist(intrafeatures[loc], intrafeatures[loc], 'cosine')
c_dists = np.sum(c_dists)/(len(c_dists)**2-len(c_dists))
intra_dists.append(c_dists)
intra_dists = np.array(intra_dists)
maxval = np.max(intra_dists[1-np.isnan(intra_dists)])
intra_dists[np.isnan(intra_dists)] = maxval
intra_dists[np.isinf(intra_dists)] = maxval
dist_metric = dist_metric_intra = np.mean(intra_dists)
if mode == 'inter':
if not isinstance(features, torch.Tensor):
coms = []
for loc in features_locs:
com = normalize(np.mean(features[loc], axis=0).reshape(1,-1)).reshape(-1)
coms.append(com)
mean_inter_dist = distance.cdist(np.array(coms), np.array(coms), 'cosine')
dist_metric = dist_metric_inter = np.sum(mean_inter_dist)/(len(mean_inter_dist)**2-len(mean_inter_dist))
else:
coms = []
for loc in features_locs:
com = torch.nn.functional.normalize(torch.mean(features[loc], dim=0).reshape(1, -1), dim=-1).reshape(1,-1)
coms.append(com)
mean_inter_dist = 1-torch.cat(coms, dim=0).mm(torch.cat(coms, dim=0).T).detach().cpu().numpy()
dist_metric = dist_metric_inter = np.sum(mean_inter_dist)/(len(mean_inter_dist)**2-len(mean_inter_dist))
if mode == 'fsu':
dist_metric = dist_metric_intra/np.clip(dist_metric_inter, 1e-8, None)
return dist_metric
parser = argparse.ArgumentParser(description='OpenMix: Exploring Outlier Samples for Misclassification Detection')
parser.add_argument('--epochs', default=200, type=int, help='Total number of epochs to run')
parser.add_argument('--batch_size', default=128, type=int, help='Batch size for training')
parser.add_argument('--plot', default=10, type=int, help='')
parser.add_argument('--run', default=3, type=int, help='')
parser.add_argument('--classnumber', default=10, type=int, help='class number for the dataset')
parser.add_argument('--classnumber_aug', default=1, type=int, help='class number for the dataset')
parser.add_argument('--data', default='cifar10', type=str, help='Dataset name to use [cifar10, cifar100]')
parser.add_argument('--model', default='res110', type=str, help='Models name to use [res110, wrn, dense, resnet18, vgg, cmixer]')
parser.add_argument('--method', default='OpenMix_manifold', type=str, help='[OpenMix, OpenMix_manifold, OpenMix-CRL, '
'classAug, Mixup, Manifold, RegMixup, '
'OE, Baseline, CRL')
parser.add_argument('--data_path', default='/data/datasets/', type=str, help='Dataset directory')
parser.add_argument('--save_path', default='./output/', type=str, help='Savefiles directory')
parser.add_argument('--rank_weight', default=1.0, type=float, help='Rank loss weight')
parser.add_argument('--gpu', default='6', type=str, help='GPU id to use')
parser.add_argument('--lambda_o', type=float, default=1, help='[0.1, 0.5, 1.0, 1.5, 2] dnl loss weight')
parser.add_argument('--print-freq', '-p', default=1, type=int, metavar='N', help='print frequency (default: 10)')
parser.add_argument('-aux_set', type=str, default='RandomImages', help='RandomImages')
parser.add_argument('-aux_size', type=int, default=-1, help='using all RandomImages data')
parser.add_argument('-prefetch', type=int, default=4, help='Pre-fetching threads.')
parser.add_argument('-aux_batch_size', type=int, default=128, help='Batch size.')
parser.add_argument('--openness', type=float, default=0.5, help='[0.1, 0.5, 1.0, 1.5, 2]')
args = parser.parse_args()
def main():
acc_list=[]
auroc_list=[]
aupr_success_list=[]
aupr_list=[]
fpr_list=[]
aurc_list=[]
eaurc_list=[]
ece_list=[]
nll_list=[]
brier_list=[]
os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu
cudnn.benchmark = True
save_path = args.save_path + args.data + '_' + args.model + '_' + args.method
if not os.path.exists(save_path):
os.makedirs(save_path)
train_loader, valid_loader, test_loader, \
test_onehot, test_label = dataset.get_loader(args.data, args.data_path, args.batch_size, args)
if args.data == 'cifar100':
num_class = 100
args.classnumber = 100
else:
args.classnumber = 10
num_class = 10
if args.method == 'OpenMix' or args.method == 'OpenMix-CRL' or args.method == 'OpenMix_manifold':
num_class = num_class + 1
if args.method == 'classAug':
if args.data == 'cifar100':
args.newclassnum = 900
num_class = num_class + args.newclassnum
else:
args.newclassnum = 45
num_class = num_class + args.newclassnum
model_dict = {
"num_classes": num_class,
}
save_path = args.save_path + args.data + '_' + args.model + '_' + args.method
save_path_model = save_path + '/'
model_state_dict = torch.load(os.path.join(save_path_model, 'model.pth'))
model.load_state_dict(model_state_dict)
model.eval()
feature_list = []
labels_list = []
for input, target, idx in test_loader:
input = input.cuda()
target = target.long().cuda()
_, feature = model(input, feature_output=True)
feature_list.append(feature.detach().cpu())
labels_list.append(target)
features = torch.cat(feature_list)
labels = torch.cat(labels_list)
dist_intra = dist_feature('intra', features, labels)
dist_inter = dist_feature('inter', features, labels)
FSU = dist_feature('fsu', features, labels)
print(dist_intra, dist_inter, FSU)
if __name__ == "__main__":
main()