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
|
"""
Date: 2022-03-08
Author: [email protected]
Desc: adversarial dataset
"""
import torch
from torch.utils.data import Dataset
import numpy as np
class AdversarialC2Data(Dataset):
"""
adversarial sample for attack
"""
def __init__(self, filename, target_class=5, keep_target=True, norm=False):
"""
:param filename:
:param target_class:
:param keep_target: Whether to keep target data
"""
data = np.load(filename)
if norm:
data[:,:-1] = (data[:,:-1] - np.min(data[:,:-1], axis=0)) / (np.max(data[:,:-1], axis=0) - np.min(data[:,:-1], axis=0)) + 1e-9
print("Adversarial Dataset Load: {}".format(filename))
if keep_target:
self.data = data
else:
self.data = np.array([x for x in data if x[-1] != target_class])
print(self.data.shape)
def __getitem__(self, index):
data = self.data[index]
sample = [x if x < 1600 else 1599 for x in data[:-1]]
sample = np.array(sample)
sample = torch.from_numpy(sample)
label = torch.from_numpy(data[-1:])
return sample, label
def __len__(self):
return self.data.shape[0]
|
