-
Notifications
You must be signed in to change notification settings - Fork 0
/
svhn.py
181 lines (150 loc) · 6.85 KB
/
svhn.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
from torchvision.datasets.vision import VisionDataset
from PIL import Image
import os
import os.path
import numpy as np
from typing import Any, Callable, Optional, Tuple
from torchvision.datasets.utils import download_url, check_integrity, verify_str_arg
from torchvision import datasets, transforms
import copy
from scipy.special import comb
from augment.cutout import Cutout
from augment.autoaugment_extra import SVHNPolicy
import torch
from torchvision.transforms import Compose, ToTensor, Normalize, Pad, RandomCrop, RandomHorizontalFlip, RandomErasing, ToPILImage
from sklearn.preprocessing import OneHotEncoder
class MY_SVHN(VisionDataset):
"""`SVHN <http://ufldl.stanford.edu/housenumbers/>`_ Dataset.
Note: The SVHN dataset assigns the label `10` to the digit `0`. However, in this Dataset,
we assign the label `0` to the digit `0` to be compatible with PyTorch loss functions which
expect the class labels to be in the range `[0, C-1]`
.. warning::
This class needs `scipy <https://docs.scipy.org/doc/>`_ to load data from `.mat` format.
Args:
root (string): Root directory of dataset where directory
``SVHN`` exists.
split (string): One of {'train', 'test', 'extra'}.
Accordingly dataset is selected. 'extra' is Extra training set.
transform (callable, optional): A function/transform that takes in an PIL image
and returns a transformed version. E.g, ``transforms.RandomCrop``
target_transform (callable, optional): A function/transform that takes in the
target and transforms it.
download (bool, optional): If true, downloads the dataset from the internet and
puts it in root directory. If dataset is already downloaded, it is not
downloaded again.
"""
split_list = {
'train': ["http://ufldl.stanford.edu/housenumbers/train_32x32.mat",
"train_32x32.mat", "e26dedcc434d2e4c54c9b2d4a06d8373"],
'test': ["http://ufldl.stanford.edu/housenumbers/test_32x32.mat",
"test_32x32.mat", "eb5a983be6a315427106f1b164d9cef3"],
'extra': ["http://ufldl.stanford.edu/housenumbers/extra_32x32.mat",
"extra_32x32.mat", "a93ce644f1a588dc4d68dda5feec44a7"]}
def __init__(
self,
root: str,
split: str = "train",
transform: Optional[Callable] = None,
target_transform: Optional[Callable] = None,
download: bool = False, rate_partial=0.3,
) -> None:
super(MY_SVHN, self).__init__(root, transform=transform,
target_transform=target_transform)
self.split = verify_str_arg(split, "split", tuple(self.split_list.keys()))
self.url = self.split_list[split][0]
self.filename = self.split_list[split][1]
self.file_md5 = self.split_list[split][2]
if download:
self.download()
if not self._check_integrity():
raise RuntimeError('Dataset not found or corrupted.' +
' You can use download=True to download it')
# import here rather than at top of file because this is
# an optional dependency for torchvision
import scipy.io as sio
# reading(loading) mat file as array
loaded_mat = sio.loadmat(os.path.join(self.root, self.filename))
self.data = loaded_mat['X']
# loading from the .mat file gives an np array of type np.uint8
# converting to np.int64, so that we have a LongTensor after
# the conversion from the numpy array
# the squeeze is needed to obtain a 1D tensor
self.labels = loaded_mat['y'].astype(np.int64).squeeze()
# the svhn dataset assigns the class label "10" to the digit 0
# this makes it inconsistent with several loss functions
# which expect the class labels to be in the range [0, C-1]
np.place(self.labels, self.labels == 10, 0)
self.data = np.transpose(self.data, (3, 2, 0, 1))
self.targets = self.labels
self.rate_partial = rate_partial
self.partial_labels = self.generate_partial_labels()
self.transform=Compose([
ToTensor(),
Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]),
])
self.transform1=Compose([
ToTensor(),
Cutout(n_holes=1, length=20),
ToPILImage(),
SVHNPolicy(),
ToTensor(),
Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]),
])
def __getitem__(self, index: int) -> Tuple[Any, Any]:
"""
Args:
index (int): Index
Returns:
tuple: (image, target) where target is index of the target class.
"""
img, target, partial_label = self.data[index], self.targets[index], self.partial_labels[index]
# doing this so that it is consistent with all other datasets
# to return a PIL Image
img = Image.fromarray(np.transpose(img, (1, 2, 0)))
if self.transform is not None:
img_ori = self.transform(img)
img1 = self.transform1(img)
img2 = self.transform1(img)
if self.target_transform is not None:
target = self.target_transform(target)
return img_ori, img1, img2, target, partial_label,index
def __len__(self) -> int:
return len(self.data)
def _check_integrity(self) -> bool:
root = self.root
md5 = self.split_list[self.split][2]
fpath = os.path.join(root, self.filename)
return check_integrity(fpath, md5)
def download(self) -> None:
md5 = self.split_list[self.split][2]
download_url(self.url, self.root, self.filename, md5)
def extra_repr(self) -> str:
return "Split: {split}".format(**self.__dict__)
def generate_partial_labels(self):
def binarize_class(y):
label = y.reshape(len(y), -1)
enc = OneHotEncoder(categories='auto')
enc.fit(label)
label = enc.transform(label).toarray().astype(np.float32)
label = torch.from_numpy(label)
return label
new_y = binarize_class(train_labels.clone())
n, c = new_y.shape[0], new_y.shape[1]
avgC = 0
new_y = binarize_class(self.targets)
n = len(self.targets)
c = max(self.targets) + 1
avgC = 0
partial_rate = self.rate_partial
print(partial_rate)
for i in range(n):
row = new_y[i, :]
row[np.where(np.random.binomial(1, partial_rate, c) == 1)] = 1
while torch.sum(row) == 1:
row[np.random.randint(0, c)] = 1
avgC += torch.sum(row)
new_y[i] = row
avgC = avgC / n
print("Finish Generating Candidate Label Sets:{}!\n".format(avgC))
new_y = new_y.cpu().numpy()
return new_y