google-research

1
# coding=utf-8
2
# Copyright 2024 The Google Research Authors.
3
#
4
# Licensed under the Apache License, Version 2.0 (the "License");
5
# you may not use this file except in compliance with the License.
6
# You may obtain a copy of the License at
7
#
8
#     http://www.apache.org/licenses/LICENSE-2.0
9
#
10
# Unless required by applicable law or agreed to in writing, software
11
# distributed under the License is distributed on an "AS IS" BASIS,
12
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
# See the License for the specific language governing permissions and
14
# limitations under the License.
15

16
# Copyright 2016 Google Inc. All Rights Reserved.
17
#
18
# Licensed under the Apache License, Version 2.0 (the "License");
19
# you may not use this file except in compliance with the License.
20
# You may obtain a copy of the License at
21
#
22
# http://www.apache.org/licenses/LICENSE-2.0
23
#
24
# Unless required by applicable law or agreed to in writing, software
25
# distributed under the License is distributed on an "AS IS" BASIS,
26
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
27
# See the License for the specific language governing permissions and
28
# limitations under the License.
29
# ==============================================================================
30

31
"""Provides data for the Cifar100 dataset."""
32

33
from __future__ import absolute_import
34
from __future__ import division
35
from __future__ import print_function
36

37
import os
38

39
import tensorflow as tf
40

41
from tensorflow.contrib.slim.python.slim.data import dataset
42
from tensorflow.contrib.slim.python.slim.data import tfexample_decoder
43

44
_FILE_PATTERN = 'cifar100_%s-*'
45

46
_DATASET_DIR = ('')
47

48
_SPLITS_TO_SIZES = {'train': 50000, 'test': 10000}
49

50
_NUM_CLASSES = 100
51

52
_ITEMS_TO_DESCRIPTIONS = {
53
    'image': 'A [32 x 32 x 3] color image.',
54
    'image/class/label': 'A single integer between 0 and 99.',
55
    'image/format': 'a string indicating the image format.',
56
    'image/class/fine_label': 'A single integer between 0 and 99.',
57
}
58

59

60
def get_split(split_name, dataset_dir=None):
61
  """Gets a dataset tuple with instructions for reading cifar100.
62

63
  Args:
64
    split_name: A train/test split name.
65
    dataset_dir: The base directory of the dataset sources.
66

67
  Returns:
68
    A `Dataset` namedtuple. Image tensors are integers in [0, 255].
69

70
  Raises:
71
    ValueError: if `split_name` is not a valid train/test split.
72
  """
73
  if split_name not in _SPLITS_TO_SIZES:
74
    raise ValueError('split name %s was not recognized.' % split_name)
75

76
  file_pattern = os.path.join(dataset_dir, _FILE_PATTERN % split_name)
77

78
  keys_to_features = {
79
      'image/encoded': tf.FixedLenFeature((), tf.string, default_value=''),
80
      'image/format': tf.FixedLenFeature((), tf.string, default_value=''),
81
      'image/class/label': tf.FixedLenFeature(
82
          [1], tf.int64, default_value=tf.zeros([1], dtype=tf.int64)),
83
      'image/class/fine_label': tf.FixedLenFeature(
84
          [1], tf.int64, default_value=tf.zeros([1], dtype=tf.int64)),
85
  }
86

87
  if split_name == 'train':
88
    items_to_handlers = {
89
        'image': tfexample_decoder.Image(shape=[32, 32, 3]),
90
        'label': tfexample_decoder.Tensor('image/class/label'),
91
    }
92
  else:
93
    items_to_handlers = {
94
        'image': tfexample_decoder.Image(shape=[32, 32, 3]),
95
        'label': tfexample_decoder.Tensor('image/class/fine_label'),
96
    }
97

98
  decoder = tfexample_decoder.TFExampleDecoder(
99
      keys_to_features, items_to_handlers)
100

101
  return dataset.Dataset(
102
      data_sources=file_pattern,
103
      reader=tf.TFRecordReader,
104
      decoder=decoder,
105
      num_samples=_SPLITS_TO_SIZES[split_name],
106
      num_classes=_NUM_CLASSES,
107
      items_to_descriptions=_ITEMS_TO_DESCRIPTIONS)
108