-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdataset.py
148 lines (132 loc) · 7.98 KB
/
dataset.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
import torch
import torchvision
from torchvision import transforms as transforms
from torch.utils.data import random_split
from lib.alibi import RandomizedLabelPrivacy
from lib.canary import fill_canaries
import numpy as np
class NoisedCIFAR(torch.utils.data.Dataset):
def __init__(
self,
cifar: torch.utils.data.Dataset,
num_classes: int,
randomized_label_privacy: RandomizedLabelPrivacy,
):
self.cifar = cifar
self.rlp = randomized_label_privacy
targets = [cifar.dataset.targets[index] for index in cifar.indices]
self.soft_targets = [self._noise(t, num_classes) for t in targets]
self.rlp.increase_budget() # increase budget
# calculate probability of label change
num_label_changes = sum(
label != torch.argmax(soft_target).item()
for label, soft_target in zip(targets, self.soft_targets)
)
self.label_change = num_label_changes / len(targets)
def _noise(self, label, n):
onehot = torch.zeros(n).to()
onehot[label] = 1
rand = self.rlp.noise((n,))
return onehot if rand is None else onehot + rand
def __len__(self):
return self.cifar.__len__()
def __getitem__(self, index):
image, label = self.cifar.__getitem__(index)
return image, self.soft_targets[index], label
def load_CIFAR10(batch_size):
CIFAR10_MEAN = (0.4914, 0.4822, 0.4465)
CIFAR10_STD_DEV = (0.2023, 0.1994, 0.2010)
train_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)]) #transforms.RandomHorizontalFlip(),
test_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)])
train_set = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=train_transform)
train_loader = torch.utils.data.DataLoader(dataset=train_set, batch_size=batch_size, shuffle=True)
test_set = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=test_transform)
test_loader = torch.utils.data.DataLoader(dataset=test_set, batch_size=batch_size, shuffle=False)
labelNames = ["airplane", "automobile", "bird", "cat", "dear", "dog", "frog", "horse", "ship", "truck"]
return train_loader, test_loader, labelNames
def load_CIFAR100(batch_size):
CIFAR100_MEAN = (0.5071, 0.4867, 0.4408)
CIFAR100_STD_DEV = (0.2675, 0.2565, 0.2761)
train_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR100_MEAN, CIFAR100_STD_DEV)]) #transforms.RandomHorizontalFlip(),
test_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR100_MEAN, CIFAR100_STD_DEV)])
train_set = torchvision.datasets.CIFAR100(root='./data', train=True, download=True, transform=train_transform)
train_loader = torch.utils.data.DataLoader(dataset=train_set, batch_size=batch_size, shuffle=True)
test_set = torchvision.datasets.CIFAR100(root='./data', train=False, download=True, transform=test_transform)
test_loader = torch.utils.data.DataLoader(dataset=test_set, batch_size=batch_size, shuffle=False)
return train_loader, test_loader
def load_FedCIFAR10(batch_size, num_workers):
CIFAR10_MEAN = (0.4914, 0.4822, 0.4465)
CIFAR10_STD_DEV = (0.2023, 0.1994, 0.2010)
train_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)]) #transforms.RandomHorizontalFlip(),
test_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)])
train_set = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=train_transform)
len_train = len(train_set)
len_client_data = {worker_id: int(len_train / num_workers) for worker_id in range(num_workers)}
if sum(len_client_data.values()) != len_train:
dif = len_train - sum(len_client_data.values())
idxs = np.random.choice(num_workers, size=dif)
for idx in idxs:
len_client_data[idx] += 1
rs = random_split(train_set, list(len_client_data.values()))
train_loaders = [(torch.utils.data.DataLoader(dataset=rs[worker_id], batch_size=batch_size, shuffle=True), len(rs[worker_id]) / len_train) for worker_id in range(num_workers)]
test_set = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=test_transform)
test_loader = torch.utils.data.DataLoader(dataset=test_set, batch_size=batch_size, shuffle=False)
labelNames = ["airplane", "automobile", "bird", "cat", "dear", "dog", "frog", "horse", "ship", "truck"]
return train_loaders, test_loader, labelNames
def load_FedCIFAR10_LAPLACE(args, randomized_label_privacy):
CIFAR10_MEAN = (0.4914, 0.4822, 0.4465)
CIFAR10_STD_DEV = (0.2023, 0.1994, 0.2010)
train_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)]) #transforms.RandomHorizontalFlip(),
test_transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize(CIFAR10_MEAN, CIFAR10_STD_DEV)])
train_set = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=train_transform)
len_train = len(train_set)
len_client_data = {worker_id: int(len_train / args.NUM_WORKERS) for worker_id in range(args.NUM_WORKERS)}
if sum(len_client_data.values()) != len_train:
dif = len_train - sum(len_client_data.values())
idxs = np.random.choice(args.NUM_WORKERS, size=dif)
for idx in idxs:
len_client_data[idx] += 1
rs = random_split(train_set, list(len_client_data.values()))
rs = [rand_label_privacy_process(ds, randomized_label_privacy, args, 10) for ds in rs]
train_loaders = [(torch.utils.data.DataLoader(dataset=rs[worker_id], batch_size=args.BATCH_SIZE, shuffle=True), len(rs[worker_id]) / len_train) for worker_id in range(args.NUM_WORKERS)]
test_set = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=test_transform)
test_loader = torch.utils.data.DataLoader(dataset=test_set, batch_size=args.BATCH_SIZE, shuffle=False)
labelNames = ["airplane", "automobile", "bird", "cat", "dear", "dog", "frog", "horse", "ship", "truck"]
return train_loaders, test_loader, labelNames
def rand_label_privacy_process(dataset, randomized_label_privacy, args, num_classes):
if args.CANARY > 0 and args.CANARY < len(dataset):
# capture debug info
original_label_sum = sum([dataset.dataset.targets[index] for index in dataset.indices])
original_last10_labels = [dataset[-i][1] for i in range(1, 11)]
# inject canaries
dataset = fill_canaries(
dataset, num_classes, N=args.CANARY, seed=args.SEED
)
# capture debug info
canary_label_sum = sum([dataset.dataset.targets[index] for index in dataset.indices])
canary_last10_labels = [dataset[-i][1] for i in range(1, 11)]
# verify presence
if original_label_sum == canary_label_sum:
raise Exception(
"Canary infiltration has failed."
f"\nOriginal label sum: {original_label_sum} vs"
f" Canary label sum: {canary_label_sum}"
f"\nOriginal last 10 labels: {original_last10_labels} vs"
f" Canary last 10 labels: {canary_last10_labels}"
)
if args.NOISE_ONLY_ONCE:
dataset = NoisedCIFAR(
dataset, num_classes, randomized_label_privacy
)
return dataset
def load_data(args, randomized_label_privacy=None):
if args.DATA_NAME == 'CIFAR10':
return load_CIFAR10(args.BATCH_SIZE)
elif args.DATA_NAME == 'CIFAR100':
return load_CIFAR100(args.BATCH_SIZE)
elif args.DATA_NAME == 'FedCIFAR10':
return load_FedCIFAR10(args.BATCH_SIZE, num_workers=args.NUM_WORKERS)
elif args.DATA_NAME == 'FedCIFAR10LAPLACE':
return load_FedCIFAR10_LAPLACE(args, randomized_label_privacy)
else:
raise ValueError('Unknown data name: {}'.format(args.DATA_NAME))