-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdata.py
332 lines (263 loc) · 11.6 KB
/
data.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
import random
import numpy as np
import torch
import torch.nn as nn
import torchvision
from torch.utils.data import Dataset, DataLoader
import torch.nn.functional as F
import torchvision
from torch.utils.data import Dataset
class MNISTInMemory(Dataset):
"""
A PyTorch dataset for loading the MNIST dataset into memory.
Args:
root (str): The root directory where the dataset will be stored. Default is "./datasets/MNIST".
train (bool): Whether to load the training or testing set. Default is True.
Attributes:
mnist_data (torchvision.datasets.MNIST): The MNIST dataset.
data (List[Tuple[torch.Tensor, int]]): The data as a list of tuples containing an image tensor and its label.
Methods:
__getitem__: Gets an item from the dataset.
__len__: Gets the length of the dataset.
"""
def __init__(self, root='./datasets/MNIST', train=True):
self.mnist_data = torchvision.datasets.MNIST(root=root, train=train, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize(
(0.1307,), (0.3081,))
]))
self.data = []
for img, target in self.mnist_data:
self.data.append((img, target))
def __getitem__(self, index):
return self.data[index]
def __len__(self):
return len(self.mnist_data)
class FMNISTInMemory(Dataset):
"""
A PyTorch dataset for loading the Fashion-MNIST dataset into memory.
Args:
root (str): The root directory where the dataset will be stored. Default is "./datasets/FMNIST".
train (bool): Whether to load the training or testing set. Default is True.
Attributes:
fmnist_data (torchvision.datasets.FashionMNIST): The Fashion-MNIST dataset.
data (List[Tuple[torch.Tensor, int]]): The data as a list of tuples containing an image tensor and its label.
Methods:
__getitem__: Gets an item from the dataset.
__len__: Gets the length of the dataset.
"""
def __init__(self, root='./datasets/FMNIST', train=True):
self.fmnist_data = torchvision.datasets.FashionMNIST(root=root, train=train, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize(
(0.28582752,), (0.35303804,))
]))
self.data = []
for img, target in self.fmnist_data:
self.data.append((img, target))
def __getitem__(self, index):
return self.data[index]
def __len__(self):
return len(self.fmnist_data)
class CIFAR10InMemory(Dataset):
"""
A PyTorch dataset for loading the CIFAR-10 dataset into memory.
Args:
root (str): The root directory where the dataset will be stored. Default is "./datasets/CIFAR10".
train (bool): Whether to load the training or testing set. Default is True.
Attributes:
cifar10_data (torchvision.datasets.CIFAR10): The CIFAR-10 dataset.
data (List[Tuple[torch.Tensor, int]]): The data as a list of tuples containing an image tensor and its label.
Methods:
__getitem__: Gets an item from the dataset.
__len__: Gets the length of the dataset.
"""
def __init__(self, root='./datasets/CIFAR10', train=True):
self.cifar10_data = torchvision.datasets.CIFAR10(root=root, train=train, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize(
(0.4914, 0.4822, 0.4465), (0.2470, 0.2435, 0.2616)) # mean and std for CIFAR-10
]))
self.data = []
for img, target in self.cifar10_data:
self.data.append((img, target))
def __getitem__(self, index):
return self.data[index]
def __len__(self):
return len(self.cifar10_data)
class MergedDataset(Dataset):
def __init__(self, original_dataset, steps, batch_size, negative_data_generator):
self.original_dataset = original_dataset
self.num_images = len(original_dataset)
self.steps = steps
self.batch_size = batch_size
self.negative_data_generator = negative_data_generator
def __getitem__(self, index):
imgs1, imgs2 = [], []
for _ in range(self.batch_size):
# Randomly select two images from the original dataset
index1 = random.randint(0, self.num_images - 1)
index2 = random.randint(0, self.num_images - 1)
img1, _ = self.original_dataset[index1]
img2, _ = self.original_dataset[index2]
imgs1.append(img1)
imgs2.append(img2)
# Stack tensors along a new dimension
imgs1 = torch.stack(imgs1)
imgs2 = torch.stack(imgs2)
# Generate negative data
neg_data = self.negative_data_generator(imgs1, imgs2, self.steps)
return neg_data
def __len__(self):
# Return the number of pairs we can make from the original dataset
return self.num_images // self.batch_size
class MNIST:
@staticmethod
def overlay_y_on_x(x: torch.Tensor, y: torch.Tensor):
_data = x.clone()
_data = _data.view(-1, _data.numel() // _data.shape[0])
overlay = torch.zeros(_data.shape[0], 10)
overlay[range(_data.shape[0]), y] = x.max()
_data[:, :10] = overlay
return _data, y
@staticmethod
def predict(data_loader: DataLoader, model: nn.Module, device="cpu"):
model.eval()
predictions = []
real = []
for i, (x, y) in enumerate(data_loader):
x = x.to(device)
y = y.to(device)
batch_size = x.shape[0]
max_goodness = torch.zeros((10, batch_size)).to(device)
for num in range(10):
over_x = MNIST.overlay_y_on_x(x.clone(), num)[0]
good = model.call(over_x)
max_goodness[num, :] = good.squeeze().detach()
_, best_nums = torch.max(max_goodness, dim=0)
predictions.extend(best_nums.cpu().numpy())
real.extend(y.cpu().numpy())
return np.array(predictions), np.array(real)
@staticmethod
def negative_data_with_masks(a, b, blurring_steps=3):
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
a = a.to(device)
b = b.to(device)
mask = torch.rand_like(a) < 0.5
mask = mask.float()
filter = torch.tensor([[1/16., 1/8., 1/16.], [1/8., 1/4., 1/8.], [1/16., 1/8., 1/16.]], device=device)
filter = filter.view(1, 1, 3, 3)
for _ in range(blurring_steps):
mask = F.conv2d(mask, filter, padding="same")
mask = mask < 0.5
mask = mask.int()
reversed_mask = torch.bitwise_not(mask.bool()).int()
neg_data = torch.mul(mask, a) + torch.mul(reversed_mask, b)
return neg_data
@staticmethod
def get_in_memory(train):
return MNISTInMemory(train = train)
@staticmethod
def get_image_shape():
return (1, 28, 28)
class FMNIST:
@staticmethod
def overlay_y_on_x(x: torch.Tensor, y: torch.Tensor):
_data = x.clone()
_data = _data.view(-1, _data.numel() // _data.shape[0])
overlay = torch.zeros(_data.shape[0], 10)
overlay[range(_data.shape[0]), y] = x.max()
_data[:, :10] = overlay
return _data, y
@staticmethod
def predict(data_loader: DataLoader, model: nn.Module, device="cpu"):
model.eval()
predictions = []
real = []
for i, (x, y) in enumerate(data_loader):
x = x.to(device)
y = y.to(device)
batch_size = x.shape[0]
max_goodness = torch.zeros((10, batch_size)).to(device)
for num in range(10):
over_x = FMNIST.overlay_y_on_x(x.clone(), num)[0]
good = model.call(over_x)
max_goodness[num, :] = good.squeeze().detach()
_, best_nums = torch.max(max_goodness, dim=0)
predictions.extend(best_nums.cpu().numpy())
real.extend(y.cpu().numpy())
return np.array(predictions), np.array(real)
@staticmethod
def get_in_memory(train):
return FMNISTInMemory(train = train)
@staticmethod
def negative_data_with_masks(a, b, blurring_steps=3):
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
a = a.to(device)
b = b.to(device)
mask = torch.rand_like(a) < 0.5
mask = mask.float()
filter = torch.tensor([[1/16., 1/8., 1/16.], [1/8., 1/4., 1/8.], [1/16., 1/8., 1/16.]], device=device)
filter = filter.view(1, 1, 3, 3)
for _ in range(blurring_steps):
mask = F.conv2d(mask, filter, padding="same")
mask = mask < 0.5
mask = mask.int()
reversed_mask = ~mask.bool()
neg_data = mask * a + reversed_mask * b
return neg_data
@staticmethod
def get_image_shape():
return (1, 28, 28)
class CIFAR10:
@staticmethod
def overlay_y_on_x(x: torch.Tensor, y: torch.Tensor):
_data = x.clone()
_data = _data.view(-1, _data.numel() // _data.shape[0])
overlay = torch.zeros(_data.shape[0], 10)
overlay[range(_data.shape[0]), y] = x.max()
_data[:, :10] = overlay
return _data, y
@staticmethod
def predict(data_loader: DataLoader, model: nn.Module, device="cpu"):
model.eval()
predictions = []
real = []
for i, (x, y) in enumerate(data_loader):
x = x.to(device)
y = y.to(device)
batch_size = x.shape[0]
max_goodness = torch.zeros((10, batch_size)).to(device)
for num in range(10):
over_x = CIFAR10.overlay_y_on_x(x.clone(), num)[0]
good = model.call(over_x)
max_goodness[num, :] = good.squeeze().detach()
_, best_nums = torch.max(max_goodness, dim=0)
predictions.extend(best_nums.cpu().numpy())
real.extend(y.cpu().numpy())
return np.array(predictions), np.array(real)
@staticmethod
def get_in_memory(train):
return CIFAR10InMemory(train = train)
@staticmethod
def negative_data_with_masks(a, b, blurring_steps=3):
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
a = a.to(device)
b = b.to(device)
mask = torch.rand_like(a) < 0.5
mask = mask.float()
filter = torch.tensor([[1/16., 1/8., 1/16.], [1/8., 1/4., 1/8.], [1/16., 1/8., 1/16.]], device=device)
filter = filter.view(1, 1, 3, 3).repeat(3, 1, 1, 1)
for _ in range(blurring_steps):
mask = F.conv2d(mask, filter, padding="same", groups=3)
mask = mask < 0.5
mask = mask.int()
reversed_mask = ~mask.bool()
neg_data = mask * a + reversed_mask * b
return neg_data
@staticmethod
def get_image_shape():
return (3, 32, 32)