forked from BoyuanJiang/matching-networks-pytorch
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdata_loader.py
190 lines (169 loc) · 9.1 KB
/
data_loader.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
# +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# Created by: BoyuanJiang
# College of Information Science & Electronic Engineering,ZheJiang University
# Email: [email protected]
# Copyright (c) 2017
# @Time :17-8-27 10:46
# @FILE :data_loader.py
# +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
import numpy as np
class OmniglotNShotDataset():
def __init__(self, batch_size, classes_per_set=20, samples_per_class=1, seed=2017, shuffle=True, use_cache=True):
"""
Construct N-shot dataset
:param batch_size: Experiment batch_size
:param classes_per_set: Integer indicating the number of classes per set
:param samples_per_class: Integer indicating samples per class
:param seed: seed for random function
:param shuffle: if shuffle the dataset
:param use_cache: if true,cache dataset to memory.It can speedup the train but require larger memory
"""
np.random.seed(seed)
self.x = np.load('data/data.npy')
self.x = np.reshape(self.x, newshape=(self.x.shape[0], self.x.shape[1], 28, 28, 1))
if shuffle:
np.random.shuffle(self.x)
self.x_train, self.x_val, self.x_test = self.x[:1200], self.x[1200:1411], self.x[1411:]
# self.mean = np.mean(list(self.x_train) + list(self.x_val))
self.x_train = self.processes_batch(self.x_train, np.mean(self.x_train), np.std(self.x_train))
self.x_test = self.processes_batch(self.x_test, np.mean(self.x_test), np.std(self.x_test))
self.x_val = self.processes_batch(self.x_val, np.mean(self.x_val), np.std(self.x_val))
# self.std = np.std(list(self.x_train) + list(self.x_val))
self.batch_size = batch_size
self.n_classes = self.x.shape[0]
self.classes_per_set = classes_per_set
self.samples_per_class = samples_per_class
self.indexes = {"train": 0, "val": 0, "test": 0}
self.datatset = {"train": self.x_train, "val": self.x_val, "test": self.x_test}
self.use_cache = use_cache
if self.use_cache:
self.cached_datatset = {"train": self.load_data_cache(self.x_train),
"val": self.load_data_cache(self.x_val),
"test": self.load_data_cache(self.x_test)}
def processes_batch(self, x_batch, mean, std):
"""
Normalizes a batch images
:param x_batch: a batch images
:return: normalized images
"""
return (x_batch - mean) / std
def _sample_new_batch(self, data_pack):
"""
Collect 1000 batches data for N-shot learning
:param data_pack: one of(train,test,val) dataset shape[classes_num,20,28,28,1]
:return: A list with [support_set_x,support_set_y,target_x,target_y] ready to be fed to our networks
"""
support_set_x = np.zeros((self.batch_size, self.classes_per_set, self.samples_per_class, data_pack.shape[2],
data_pack.shape[3], data_pack.shape[4]), np.float32)
support_set_y = np.zeros((self.batch_size, self.classes_per_set, self.samples_per_class), np.int32)
target_x = np.zeros((self.batch_size, data_pack.shape[2], data_pack.shape[3], data_pack.shape[4]), np.float32)
target_y = np.zeros((self.batch_size, 1), np.int32)
for i in range(self.batch_size):
classes_idx = np.arange(data_pack.shape[0])
samples_idx = np.arange(data_pack.shape[1])
choose_classes = np.random.choice(classes_idx, size=self.classes_per_set, replace=False)
choose_label = np.random.choice(self.classes_per_set, size=1)
choose_samples = np.random.choice(samples_idx, size=self.samples_per_class + 1, replace=False)
x_temp = data_pack[choose_classes]
x_temp = x_temp[:, choose_samples]
y_temp = np.arange(self.classes_per_set)
support_set_x[i] = x_temp[:, :-1]
support_set_y[i] = np.expand_dims(y_temp[:], axis=1)
target_x[i] = x_temp[choose_label, -1]
target_y[i] = y_temp[choose_label]
return support_set_x, support_set_y, target_x, target_y
def _rotate_data(self, image, k):
"""
Rotates one image by self.k * 90 degrees counter-clockwise
:param image: Image to rotate
:return: Rotated Image
"""
return np.rot90(image, k)
def _rotate_batch(self, batch_images, k):
"""
Rotates a whole image batch
:param batch_images: A batch of images
:param k: integer degree of rotation counter-clockwise
:return: The rotated batch of images
"""
batch_size = batch_images.shape[0]
for i in np.arange(batch_size):
batch_images[i] = self._rotate_data(batch_images[i], k)
return batch_images
def _get_batch(self, dataset_name, augment=False):
"""
Get next batch from the dataset with name.
:param dataset_name: The name of dataset(one of "train","val","test")
:param augment: if rotate the images
:return: a batch images
"""
if self.use_cache:
support_set_x, support_set_y, target_x, target_y = self._get_batch_from_cache(dataset_name)
else:
support_set_x, support_set_y, target_x, target_y = self._sample_new_batch(self.datatset[dataset_name])
if augment:
k = np.random.randint(0, 4, size=(self.batch_size, self.classes_per_set))
a_support_set_x = []
a_target_x = []
for b in range(self.batch_size):
temp_class_set = []
for c in range(self.classes_per_set):
temp_class_set_x = self._rotate_batch(support_set_x[b, c], k=k[b, c])
if target_y[b] == support_set_y[b, c, 0]:
temp_target_x = self._rotate_data(target_x[b], k=k[b, c])
temp_class_set.append(temp_class_set_x)
a_support_set_x.append(temp_class_set)
a_target_x.append(temp_target_x)
support_set_x = np.array(a_support_set_x)
target_x = np.array(a_target_x)
support_set_x = support_set_x.reshape((support_set_x.shape[0], support_set_x.shape[1] * support_set_x.shape[2],
support_set_x.shape[3], support_set_x.shape[4], support_set_x.shape[5]))
support_set_y = support_set_y.reshape(support_set_y.shape[0], support_set_y.shape[1] * support_set_y.shape[2])
return support_set_x, support_set_y, target_x, target_y
def get_train_batch(self, augment=False):
return self._get_batch("train", augment)
def get_val_batch(self, augment=False):
return self._get_batch("val", augment)
def get_test_batch(self, augment=False):
return self._get_batch("test", augment)
def load_data_cache(self, data_pack, argument=True):
"""
cache the dataset in memory
:param data_pack: shape[classes_num,20,28,28,1]
:return:
"""
cached_dataset = []
classes_idx = np.arange(data_pack.shape[0])
samples_idx = np.arange(data_pack.shape[1])
for _ in range(1000):
support_set_x = np.zeros((self.batch_size, self.classes_per_set, self.samples_per_class, data_pack.shape[2],
data_pack.shape[3], data_pack.shape[4]), np.float32)
support_set_y = np.zeros((self.batch_size, self.classes_per_set, self.samples_per_class), np.int32)
target_x = np.zeros((self.batch_size, data_pack.shape[2], data_pack.shape[3], data_pack.shape[4]),
np.float32)
target_y = np.zeros((self.batch_size, 1), np.int32)
for i in range(self.batch_size):
choose_classes = np.random.choice(classes_idx, size=self.classes_per_set, replace=False)
choose_label = np.random.choice(self.classes_per_set, size=1)
choose_samples = np.random.choice(samples_idx, size=self.samples_per_class + 1, replace=False)
x_temp = data_pack[choose_classes]
x_temp = x_temp[:, choose_samples]
y_temp = np.arange(self.classes_per_set)
support_set_x[i] = x_temp[:, :-1]
support_set_y[i] = np.expand_dims(y_temp[:], axis=1)
target_x[i] = x_temp[choose_label, -1]
target_y[i] = y_temp[choose_label]
cached_dataset.append([support_set_x, support_set_y, target_x, target_y])
return cached_dataset
def _get_batch_from_cache(self, dataset_name):
"""
:param dataset_name:
:return:
"""
if self.indexes[dataset_name] >= len(self.cached_datatset[dataset_name]):
self.indexes[dataset_name] = 0
self.cached_datatset[dataset_name] = self.load_data_cache(self.datatset[dataset_name])
next_batch = self.cached_datatset[dataset_name][self.indexes[dataset_name]]
self.indexes[dataset_name] += 1
x_support_set, y_support_set, x_target, y_target = next_batch
return x_support_set, y_support_set, x_target, y_target