cifar.py 2.5 KB
Newer Older
D
dangqingqing 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13
# Copyright (c) 2016 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.
Y
Yu Yang 已提交
14
"""
Y
Yi Wang 已提交
15
CIFAR dataset: https://www.cs.toronto.edu/~kriz/cifar.html
Y
Yu Yang 已提交
16 17

TODO(yuyang18): Complete the comments.
Y
Yu Yang 已提交
18
"""
D
dangqingqing 已提交
19

Y
Yu Yang 已提交
20 21 22
import cPickle
import itertools
import numpy
Y
Yi Wang 已提交
23 24
import paddle.v2.dataset.common
import tarfile
Y
Yu Yang 已提交
25

Y
Yi Wang 已提交
26
__all__ = ['train100', 'test100', 'train10', 'test10']
Y
Yu Yang 已提交
27

Y
Yi Wang 已提交
28 29
URL_PREFIX = 'https://www.cs.toronto.edu/~kriz/'
CIFAR10_URL = URL_PREFIX + 'cifar-10-python.tar.gz'
Y
Yu Yang 已提交
30
CIFAR10_MD5 = 'c58f30108f718f92721af3b95e74349a'
Y
Yi Wang 已提交
31
CIFAR100_URL = URL_PREFIX + 'cifar-100-python.tar.gz'
Y
Yu Yang 已提交
32 33 34
CIFAR100_MD5 = 'eb9058c3a382ffc7106e4002c42a8d85'


Y
Yi Wang 已提交
35 36 37 38 39 40 41
def reader_creator(filename, sub_name):
    def read_batch(batch):
        data = batch['data']
        labels = batch.get('labels', batch.get('fine_labels', None))
        assert labels is not None
        for sample, label in itertools.izip(data, labels):
            yield (sample / 255.0).astype(numpy.float32), int(label)
Y
Yu Yang 已提交
42

Y
Yi Wang 已提交
43
    def reader():
Y
Yu Yang 已提交
44
        with tarfile.open(filename, mode='r') as f:
Y
Yu Yang 已提交
45 46 47 48 49
            names = (each_item.name for each_item in f
                     if sub_name in each_item.name)

            for name in names:
                batch = cPickle.load(f.extractfile(name))
Y
Yi Wang 已提交
50
                for item in read_batch(batch):
Y
Yu Yang 已提交
51 52
                    yield item

Y
Yu Yang 已提交
53
    return reader
Y
Yu Yang 已提交
54 55


Y
Yi Wang 已提交
56 57 58 59
def train100():
    return reader_creator(
        paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5),
        'train')
Y
Yu Yang 已提交
60 61


Y
Yi Wang 已提交
62 63 64 65
def test100():
    return reader_creator(
        paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5),
        'test')
Y
Yu Yang 已提交
66 67


Y
Yi Wang 已提交
68 69 70 71
def train10():
    return reader_creator(
        paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5),
        'data_batch')
Y
Yu Yang 已提交
72 73


Y
Yi Wang 已提交
74 75 76 77
def test10():
    return reader_creator(
        paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5),
        'test_batch')
Y
Yancey1989 已提交
78 79


Y
update  
Yancey1989 已提交
80
def fetch_data():
Y
Yancey1989 已提交
81 82
    paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5)
    paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5)