-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdataloader_cifar_idn.py
395 lines (356 loc) · 16.9 KB
/
dataloader_cifar_idn.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
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
import copy
from torch.utils.data import Dataset, DataLoader
import torchvision.transforms as transforms
import random
import numpy as np
from PIL import Image
import json
import os
import torch
from torchnet.meter import AUCMeter
from autoaugment import CIFAR10Policy
def unpickle(file):
import _pickle as cPickle
with open(file, 'rb') as fo:
dict = cPickle.load(fo, encoding='latin1')
return dict
class cifar_dataset(Dataset):
def __init__(self, dataset, r, noise_mode, root_dir, transform, mode, noise_file='', pred=[], probability=[],
log='', second_ind=False):
self.r = r # noise ratio
self.transform = transform
self.mode = mode
self.transition = {0: 0, 2: 0, 4: 7, 7: 7, 1: 1, 9: 1, 3: 5, 5: 3, 6: 6,
8: 8} # class transition for asymmetric noise
self.second_ind = second_ind
self.num_samples = 50000
self.num_classes = 10 if dataset == 'cifar10' else 100
if self.mode == 'test':
if dataset == 'cifar10':
test_dic = unpickle('%s/test_batch' % root_dir)
self.test_data = test_dic['data']
self.test_data = self.test_data.reshape((10000, 3, 32, 32))
self.test_data = self.test_data.transpose((0, 2, 3, 1))
self.test_label = test_dic['labels']
elif dataset == 'cifar100':
test_dic = unpickle('%s/test' % root_dir)
self.test_data = test_dic['data']
self.test_data = self.test_data.reshape((10000, 3, 32, 32))
self.test_data = self.test_data.transpose((0, 2, 3, 1))
self.test_label = test_dic['fine_labels']
else:
train_data = []
train_label = []
if dataset == 'cifar10':
for n in range(1, 6):
dpath = '%s/data_batch_%d' % (root_dir, n)
data_dic = unpickle(dpath)
train_data.append(data_dic['data'])
train_label = train_label + data_dic['labels']
train_data = np.concatenate(train_data)
elif dataset == 'cifar100':
train_dic = unpickle('%s/train' % root_dir)
train_data = train_dic['data']
train_label = train_dic['fine_labels']
train_data = train_data.reshape((50000, 3, 32, 32))
train_data = train_data.transpose((0, 2, 3, 1))
self.temp_targets = np.array(train_label)
self.temp_data = np.array(train_data)
self.min_target = min(self.temp_targets)
self.max_target = max(self.temp_targets)
noise_types = "plc"
if os.path.exists(noise_file):
if noise_types == "plc":
noise_label1 = json.load(open(noise_file, "r"))['noise_labels']
noise_label = copy.deepcopy(train_label)
noise_label[:len(noise_label1)] = noise_label1
noise_num = len(np.where(np.array(noise_label) == np.array(train_label))[0])
print('plc idn noisy file found, %d.\n'%noise_num)
else:
# part-dependent label noise, xia xiaobo
noise_label = json.load(open(noise_file, "r"))['noise_labels']
noise_num = len(np.where(np.array(noise_label) == np.array(train_label))[0])
print('pdl idn noisy file found, %d.\n'%noise_num)
else:
if noise_types == "plc":
print('error file')# inject noise
assert 1
else:
noise_label = self.instance_noise(tau=r)
noise_label = np.array(noise_label).astype(np.int64)
print("Save pdl idn noisy labels to %s ..." % noise_file)
np.savez(noise_file, noise_labels=noise_label)
noise_label = np.array(noise_label).astype(np.int64)
train_label = np.array(train_label).astype(np.int64)
if self.mode == 'all':
self.train_data = train_data
self.noise_label = np.array(noise_label).astype(np.int64)
self.train_label = np.array(train_label).astype(np.int64)
else:
if self.mode == "labeled":
pred_idx = pred.nonzero()[0]
self.probability = [probability[i] for i in pred_idx]
clean = (np.array(noise_label) == np.array(train_label))
auc_meter = AUCMeter()
auc_meter.reset()
auc_meter.add(probability, clean)
auc, _, _ = auc_meter.value()
log.write('Numer of labeled samples:%d AUC:%.3f\n' % (pred.sum(), auc))
log.flush()
elif self.mode == "unlabeled":
pred_idx = (1 - pred).nonzero()[0]
self.train_data = train_data[pred_idx]
# self.noise_label = [noise_label[i] for i in pred_idx]
# pred_idx = np.array(pred_idx)
self.noise_label = np.array(noise_label)[pred_idx]
print("%s data has a size of %d" % (self.mode, len(self.noise_label)))
def instance_noise(
self,
tau: float = 0.2,
std: float = 0.1,
feature_size: int = 3 * 32 * 32,
# seed: int = 1
):
'''
Thanks the code from https://github.com/SML-Group/Label-Noise-Learning wrote by SML-Group.
LabNoise referred much about the generation of instance-dependent label noise from this repo.
'''
from scipy import stats
from math import inf
import torch.nn.functional as F
# np.random.seed(int(seed))
# torch.manual_seed(int(seed))
# torch.cuda.manual_seed(int(seed))
# common-used parameters
num_samples = self.num_samples
num_classes = self.num_classes
P = []
# sample instance flip rates q from the truncated normal distribution N(\tau, {0.1}^2, [0, 1])
flip_distribution = stats.truncnorm((0 - tau) / std, (1 - tau) / std,
loc=tau,
scale=std)
'''
The standard form of this distribution is a standard normal truncated to the range [a, b]
notice that a and b are defined over the domain of the standard normal.
To convert clip values for a specific mean and standard deviation, use:
a, b = (myclip_a - my_mean) / my_std, (myclip_b - my_mean) / my_std
truncnorm takes and as shape parameters.
so the above `flip_distribution' give a truncated standard normal distribution with mean = `tau`,
range = [0, 1], std = `std`
'''
# import ipdb; ipdb.set_trace()
# how many random variates you need to get
q = flip_distribution.rvs(num_samples)
# sample W \in \mathcal{R}^{S \times K} from the standard normal distribution N(0, 1^2)
W = torch.tensor(
np.random.randn(num_classes, feature_size,
num_classes)).float().cuda() #K*dim*K, dim=3072
for i in range(num_samples):
x, y = self.transform(Image.fromarray(self.temp_data[i])), torch.tensor(self.temp_targets[i])
x = x.cuda()
# step (4). generate instance-dependent flip rates
# 1 x feature_size * feature_size x 10 = 1 x 10, p is a 1 x 10 vector
p = x.reshape(1, -1).mm(W[y]).squeeze(0) #classes
# step (5). control the diagonal entry of the instance-dependent transition matrix
# As exp^{-inf} = 0, p_{y} will be 0 after softmax function.
p[y] = -inf
# step (6). make the sum of the off-diagonal entries of the y_i-th row to be q_i
p = q[i] * F.softmax(p, dim=0)
p[y] += 1 - q[i]
P.append(p)
P = torch.stack(P, 0).cpu().numpy()
l = [i for i in range(self.min_target, self.max_target + 1)]
new_label = [np.random.choice(l, p=P[i]) for i in range(num_samples)]
print('noise rate = ', (new_label != np.array(self.temp_targets)).mean())
new_targets = new_label
new_targets = np.array(new_targets).astype(np.int64)
return new_targets
def __getitem__(self, index):
if self.second_ind:
if self.mode == 'labeled':
img, target, prob = self.train_data[index], self.noise_label[index], self.probability[index]
img = Image.fromarray(img)
img1 = self.transform[0](img)
img2 = self.transform[1](img)
img3 = self.transform[2](img)
img4 = self.transform[3](img)
return img1, img2, img3, img4, target, prob
elif self.mode == 'unlabeled':
img = self.train_data[index]
img = Image.fromarray(img)
img1 = self.transform[0](img)
img2 = self.transform[1](img)
img3 = self.transform[2](img)
img4 = self.transform[3](img)
return img1, img2, img3, img4
else:
if self.mode == 'labeled':
img, target, prob = self.train_data[index], self.noise_label[index], self.probability[index]
img = Image.fromarray(img)
img1 = self.transform(img)
img2 = self.transform(img)
return img1, img2, target, prob
elif self.mode == 'unlabeled':
img = self.train_data[index]
img = Image.fromarray(img)
img1 = self.transform(img)
img2 = self.transform(img)
return img1, img2
if self.mode == 'all':
img, target = self.train_data[index], self.noise_label[index]
img = Image.fromarray(img)
img = self.transform(img)
return img, target, index
elif self.mode == 'test':
img, target = self.test_data[index], self.test_label[index]
img = Image.fromarray(img)
img = self.transform(img)
return img, target
def __len__(self):
if self.mode != 'test':
return len(self.train_data)
else:
return len(self.test_data)
class cifar_dataloader():
def __init__(self, dataset, r, noise_mode, batch_size, num_workers, root_dir, log, noise_file=''):
self.dataset = dataset
self.r = r
self.noise_mode = noise_mode
self.batch_size = batch_size
self.num_workers = num_workers
self.root_dir = root_dir
self.log = log
self.noise_file = noise_file
if self.dataset == 'cifar10':
transform_weak_10 = transforms.Compose(
[
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),
]
)
transform_strong_10 = transforms.Compose(
[
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
CIFAR10Policy(),
transforms.ToTensor(),
transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),
]
)
self.transforms = {
"warmup": transform_weak_10,
"unlabeled": [
transform_weak_10,
transform_weak_10,
transform_strong_10,
transform_strong_10
],
"labeled": [
transform_weak_10,
transform_weak_10,
transform_strong_10,
transform_strong_10
],
}
self.transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),
])
elif self.dataset == 'cifar100':
transform_weak_100 = transforms.Compose(
[
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize((0.507, 0.487, 0.441), (0.267, 0.256, 0.276)),
]
)
transform_strong_100 = transforms.Compose(
[
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
CIFAR10Policy(),
transforms.ToTensor(),
transforms.Normalize((0.507, 0.487, 0.441), (0.267, 0.256, 0.276)),
]
)
self.transforms = {
"warmup": transform_weak_100,
"unlabeled": [
transform_weak_100,
transform_weak_100,
transform_strong_100,
transform_strong_100
],
"labeled": [
transform_weak_100,
transform_weak_100,
transform_strong_100,
transform_strong_100
],
}
self.transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.507, 0.487, 0.441), (0.267, 0.256, 0.276)),
])
self.transform_train = self.transforms
def run(self, mode, pred=[], prob=[], second_ind=False):
if mode == 'warmup':
all_dataset = cifar_dataset(dataset=self.dataset, noise_mode=self.noise_mode, r=self.r,
root_dir=self.root_dir, transform=self.transform_train["warmup"], mode="all",
noise_file=self.noise_file, second_ind=second_ind)
trainloader = DataLoader(
dataset=all_dataset,
batch_size=self.batch_size * 2,
shuffle=True,
num_workers=self.num_workers)
return trainloader
elif mode == 'train':
labeled_dataset = cifar_dataset(dataset=self.dataset, noise_mode=self.noise_mode, r=self.r,
root_dir=self.root_dir, transform=self.transform_train["labeled"],
mode="labeled",
noise_file=self.noise_file, pred=pred, probability=prob, log=self.log,
second_ind=second_ind)
labeled_trainloader = DataLoader(
dataset=labeled_dataset,
batch_size=self.batch_size,
shuffle=True,
num_workers=self.num_workers)
unlabeled_dataset = cifar_dataset(dataset=self.dataset, noise_mode=self.noise_mode, r=self.r,
root_dir=self.root_dir, transform=self.transform_train["unlabeled"],
mode="unlabeled",
noise_file=self.noise_file, pred=pred, second_ind=second_ind)
unlabeled_trainloader = DataLoader(
dataset=unlabeled_dataset,
batch_size=self.batch_size,
shuffle=True,
num_workers=self.num_workers)
unlabeled_trainloader_map = DataLoader(
dataset=unlabeled_dataset,
batch_size=self.batch_size,
shuffle=False,
num_workers=self.num_workers)
return labeled_trainloader, unlabeled_trainloader, unlabeled_trainloader_map
elif mode == 'test':
test_dataset = cifar_dataset(dataset=self.dataset, noise_mode=self.noise_mode, r=self.r,
root_dir=self.root_dir, transform=self.transform_test, mode='test',
second_ind=second_ind)
test_loader = DataLoader(
dataset=test_dataset,
batch_size=self.batch_size * 2,
shuffle=False,
num_workers=self.num_workers)
return test_loader
elif mode == 'eval_train':
eval_dataset = cifar_dataset(dataset=self.dataset, noise_mode=self.noise_mode, r=self.r,
root_dir=self.root_dir,
transform=self.transform_test, mode='all', noise_file=self.noise_file,
second_ind=second_ind)
eval_loader = DataLoader(
dataset=eval_dataset,
batch_size=self.batch_size * 2,
shuffle=False,
num_workers=self.num_workers)
return eval_loader