test_datasets.py 8.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import unittest
import os
import numpy as np
import tempfile
import shutil
import cv2

22
import paddle.vision.transforms as T
23
from paddle.vision.datasets import DatasetFolder, ImageFolder, MNIST, FashionMNIST, Flowers
24
from paddle.dataset.common import _check_exists_and_download
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


class TestFolderDatasets(unittest.TestCase):
    def setUp(self):
        self.data_dir = tempfile.mkdtemp()
        self.empty_dir = tempfile.mkdtemp()
        for i in range(2):
            sub_dir = os.path.join(self.data_dir, 'class_' + str(i))
            if not os.path.exists(sub_dir):
                os.makedirs(sub_dir)
            for j in range(2):
                fake_img = (np.random.random((32, 32, 3)) * 255).astype('uint8')
                cv2.imwrite(os.path.join(sub_dir, str(j) + '.jpg'), fake_img)

    def tearDown(self):
        shutil.rmtree(self.data_dir)

    def test_dataset(self):
        dataset_folder = DatasetFolder(self.data_dir)

        for _ in dataset_folder:
            pass

        assert len(dataset_folder) == 4
        assert len(dataset_folder.classes) == 2

        dataset_folder = DatasetFolder(self.data_dir)
        for _ in dataset_folder:
            pass

    def test_folder(self):
        loader = ImageFolder(self.data_dir)

        for _ in loader:
            pass

        loader = ImageFolder(self.data_dir)
        for _ in loader:
            pass

        assert len(loader) == 4

    def test_transform(self):
        def fake_transform(img):
            return img

        transfrom = fake_transform
        dataset_folder = DatasetFolder(self.data_dir, transform=transfrom)

        for _ in dataset_folder:
            pass

        loader = ImageFolder(self.data_dir, transform=transfrom)
        for _ in loader:
            pass

    def test_errors(self):
        with self.assertRaises(RuntimeError):
            ImageFolder(self.empty_dir)
        with self.assertRaises(RuntimeError):
            DatasetFolder(self.empty_dir)

        with self.assertRaises(ValueError):
            _check_exists_and_download('temp_paddle', None, None, None, False)


class TestMNISTTest(unittest.TestCase):
    def test_main(self):
93 94
        transform = T.Transpose()
        mnist = MNIST(mode='test', transform=transform)
95 96
        self.assertTrue(len(mnist) == 10000)

97 98 99 100 101 102 103
        i = np.random.randint(0, len(mnist) - 1)
        image, label = mnist[i]
        self.assertTrue(image.shape[0] == 1)
        self.assertTrue(image.shape[1] == 28)
        self.assertTrue(image.shape[2] == 28)
        self.assertTrue(label.shape[0] == 1)
        self.assertTrue(0 <= int(label) <= 9)
104 105 106 107


class TestMNISTTrain(unittest.TestCase):
    def test_main(self):
108 109
        transform = T.Transpose()
        mnist = MNIST(mode='train', transform=transform)
110 111
        self.assertTrue(len(mnist) == 60000)

112 113 114 115 116 117 118
        i = np.random.randint(0, len(mnist) - 1)
        image, label = mnist[i]
        self.assertTrue(image.shape[0] == 1)
        self.assertTrue(image.shape[1] == 28)
        self.assertTrue(image.shape[2] == 28)
        self.assertTrue(label.shape[0] == 1)
        self.assertTrue(0 <= int(label) <= 9)
119

120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
        # test cv2 backend
        mnist = MNIST(mode='train', transform=transform, backend='cv2')
        self.assertTrue(len(mnist) == 60000)

        for i in range(len(mnist)):
            image, label = mnist[i]
            self.assertTrue(image.shape[0] == 1)
            self.assertTrue(image.shape[1] == 28)
            self.assertTrue(image.shape[2] == 28)
            self.assertTrue(label.shape[0] == 1)
            self.assertTrue(0 <= int(label) <= 9)
            break

        with self.assertRaises(ValueError):
            mnist = MNIST(mode='train', transform=transform, backend=1)

136

L
LielinJiang 已提交
137 138 139 140 141 142
class TestFASHIONMNISTTest(unittest.TestCase):
    def test_main(self):
        transform = T.Transpose()
        mnist = FashionMNIST(mode='test', transform=transform)
        self.assertTrue(len(mnist) == 10000)

143 144 145 146 147 148 149
        i = np.random.randint(0, len(mnist) - 1)
        image, label = mnist[i]
        self.assertTrue(image.shape[0] == 1)
        self.assertTrue(image.shape[1] == 28)
        self.assertTrue(image.shape[2] == 28)
        self.assertTrue(label.shape[0] == 1)
        self.assertTrue(0 <= int(label) <= 9)
L
LielinJiang 已提交
150 151 152 153 154 155 156 157


class TestFASHIONMNISTTrain(unittest.TestCase):
    def test_main(self):
        transform = T.Transpose()
        mnist = FashionMNIST(mode='train', transform=transform)
        self.assertTrue(len(mnist) == 60000)

158 159 160 161 162 163 164
        i = np.random.randint(0, len(mnist) - 1)
        image, label = mnist[i]
        self.assertTrue(image.shape[0] == 1)
        self.assertTrue(image.shape[1] == 28)
        self.assertTrue(image.shape[2] == 28)
        self.assertTrue(label.shape[0] == 1)
        self.assertTrue(0 <= int(label) <= 9)
L
LielinJiang 已提交
165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181

        # test cv2 backend
        mnist = FashionMNIST(mode='train', transform=transform, backend='cv2')
        self.assertTrue(len(mnist) == 60000)

        for i in range(len(mnist)):
            image, label = mnist[i]
            self.assertTrue(image.shape[0] == 1)
            self.assertTrue(image.shape[1] == 28)
            self.assertTrue(image.shape[2] == 28)
            self.assertTrue(label.shape[0] == 1)
            self.assertTrue(0 <= int(label) <= 9)
            break

        with self.assertRaises(ValueError):
            mnist = FashionMNIST(mode='train', transform=transform, backend=1)

L
LielinJiang 已提交
182 183 184 185 186 187 188
    def test_dataset_value(self):
        fmnist = FashionMNIST(mode='train')
        value = np.mean([np.array(x[0]) for x in fmnist])

        # 72.94035223214286 was getted from competitive products
        np.testing.assert_allclose(value, 72.94035223214286)

L
LielinJiang 已提交
189

190 191 192 193 194 195 196 197 198
class TestFlowersTrain(unittest.TestCase):
    def test_main(self):
        flowers = Flowers(mode='train')
        self.assertTrue(len(flowers) == 6149)

        # traversal whole dataset may cost a
        # long time, randomly check 1 sample
        idx = np.random.randint(0, 6149)
        image, label = flowers[idx]
199
        image = np.array(image)
200 201 202 203 204 205 206 207 208 209 210 211 212 213
        self.assertTrue(len(image.shape) == 3)
        self.assertTrue(image.shape[2] == 3)
        self.assertTrue(label.shape[0] == 1)


class TestFlowersValid(unittest.TestCase):
    def test_main(self):
        flowers = Flowers(mode='valid')
        self.assertTrue(len(flowers) == 1020)

        # traversal whole dataset may cost a
        # long time, randomly check 1 sample
        idx = np.random.randint(0, 1020)
        image, label = flowers[idx]
214
        image = np.array(image)
215 216 217 218 219 220 221 222 223 224 225 226 227 228
        self.assertTrue(len(image.shape) == 3)
        self.assertTrue(image.shape[2] == 3)
        self.assertTrue(label.shape[0] == 1)


class TestFlowersTest(unittest.TestCase):
    def test_main(self):
        flowers = Flowers(mode='test')
        self.assertTrue(len(flowers) == 1020)

        # traversal whole dataset may cost a
        # long time, randomly check 1 sample
        idx = np.random.randint(0, 1020)
        image, label = flowers[idx]
229
        image = np.array(image)
230 231 232 233
        self.assertTrue(len(image.shape) == 3)
        self.assertTrue(image.shape[2] == 3)
        self.assertTrue(label.shape[0] == 1)

234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
        # test cv2 backend
        flowers = Flowers(mode='test', backend='cv2')
        self.assertTrue(len(flowers) == 1020)

        # traversal whole dataset may cost a
        # long time, randomly check 1 sample
        idx = np.random.randint(0, 1020)
        image, label = flowers[idx]

        self.assertTrue(len(image.shape) == 3)
        self.assertTrue(image.shape[2] == 3)
        self.assertTrue(label.shape[0] == 1)

        with self.assertRaises(ValueError):
            flowers = Flowers(mode='test', backend=1)

250 251 252

if __name__ == '__main__':
    unittest.main()