-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdata_utils.py
149 lines (127 loc) · 5.38 KB
/
data_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
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.nn.parallel
import torch.backends.cudnn as cudnn
import torch.optim
import torch.utils.data
import torchvision.transforms as transforms
import torchvision
import numpy as np
import copy
np.random.seed(6)
#random.seed(2)
def build_dataset(dataset,num_meta):
normalize = transforms.Normalize(mean=[x / 255.0 for x in [125.3, 123.0, 113.9]],
std=[x / 255.0 for x in [63.0, 62.1, 66.7]])
transform_train = transforms.Compose([
transforms.ToTensor(),
transforms.Lambda(lambda x: F.pad(x.unsqueeze(0),
(4, 4, 4, 4), mode='reflect').squeeze()),
transforms.ToPILImage(),
transforms.RandomCrop(32),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
normalize,
])
transform_train_1 = transforms.Compose([
transforms.ToTensor(),
transforms.Lambda(lambda x: F.pad(x.unsqueeze(0),
(4, 4, 4, 4), mode='reflect').squeeze()),
transforms.ToPILImage(),
transforms.RandomCrop(32),
])
transform_test = transforms.Compose([
transforms.ToTensor(),
normalize
])
if dataset == 'cifar10':
train_dataset = torchvision.datasets.CIFAR10(root='../data', train=True, download=True, transform=transform_train)
train_dataset_ori = torchvision.datasets.CIFAR10(root='../data', train=True, download=False, transform=transform_train_1)
test_dataset = torchvision.datasets.CIFAR10('../data', train=False, transform=transform_test)
img_num_list = [num_meta] * 10
num_classes = 10
if dataset == 'cifar100':
train_dataset = torchvision.datasets.CIFAR100(root='../data', train=True, download=True, transform=transform_train)
train_dataset_ori = torchvision.datasets.CIFAR100(root='../data', train=True, download=False, transform=transform_train_1)
test_dataset = torchvision.datasets.CIFAR100('../data', train=False, transform=transform_test)
img_num_list = [num_meta] * 100
num_classes = 100
data_list_val = {}
for j in range(num_classes):
data_list_val[j] = [i for i, label in enumerate(train_dataset.targets) if label == j]
idx_to_meta = []
idx_to_train = []
print(img_num_list)
for cls_idx, img_id_list in data_list_val.items():
np.random.shuffle(img_id_list)
img_num = img_num_list[int(cls_idx)]
idx_to_meta.extend(img_id_list[:img_num])
idx_to_train.extend(img_id_list[img_num:])
train_data = copy.deepcopy(train_dataset)
train_data_meta = copy.deepcopy(train_dataset)
train_data_ori = copy.deepcopy(train_dataset_ori)
train_data_meta.data = np.delete(train_dataset.data,idx_to_train,axis=0)
train_data_meta.targets = np.delete(train_dataset.targets, idx_to_train, axis=0)
train_data.data = np.delete(train_dataset.data, idx_to_meta, axis=0)
train_data.targets = np.delete(train_dataset.targets, idx_to_meta, axis=0)
train_data_ori.data = np.delete(train_dataset_ori.data, idx_to_meta, axis=0)
train_data_ori.targets = np.delete(train_dataset_ori.targets, idx_to_meta, axis=0)
return train_data_meta,train_data, train_data_ori, test_dataset
def get_img_num_per_cls(dataset,imb_factor=None,num_meta=None):
"""
Get a list of image numbers for each class, given cifar version
Num of imgs follows emponential distribution
img max: 5000 / 500 * e^(-lambda * 0);
img min: 5000 / 500 * e^(-lambda * int(cifar_version - 1))
exp(-lambda * (int(cifar_version) - 1)) = img_max / img_min
args:
cifar_version: str, '10', '100', '20'
imb_factor: float, imbalance factor: img_min/img_max,
None if geting default cifar data number
output:
img_num_per_cls: a list of number of images per class
"""
if dataset == 'cifar10':
img_max = (50000-num_meta)/10
cls_num = 10
if dataset == 'cifar100':
img_max = (50000-num_meta)/100
cls_num = 100
if imb_factor is None:
return [img_max] * cls_num
img_num_per_cls = []
for cls_idx in range(cls_num):
num = img_max * (imb_factor**(cls_idx / (cls_num - 1.0)))
img_num_per_cls.append(int(num))
return img_num_per_cls
# This function is used to generate imbalanced test set
'''
def get_img_num_per_cls_test(dataset,imb_factor=None,num_meta=None):
"""
Get a list of image numbers for each class, given cifar version
Num of imgs follows emponential distribution
img max: 5000 / 500 * e^(-lambda * 0);
img min: 5000 / 500 * e^(-lambda * int(cifar_version - 1))
exp(-lambda * (int(cifar_version) - 1)) = img_max / img_min
args:
cifar_version: str, '10', '100', '20'
imb_factor: float, imbalance factor: img_min/img_max,
None if geting default cifar data number
output:
img_num_per_cls: a list of number of images per class
"""
if dataset == 'cifar10':
img_max = (10000-num_meta)/10
cls_num = 10
if dataset == 'cifar100':
img_max = (10000-num_meta)/100
cls_num = 100
if imb_factor is None:
return [img_max] * cls_num
img_num_per_cls = []
for cls_idx in range(cls_num):
num = img_max * (imb_factor**(cls_idx / (cls_num - 1.0)))
img_num_per_cls.append(int(num))
return img_num_per_cls
'''