ydshieh commited on
Commit
3fbabfa
1 Parent(s): e1e7da9
coco_dataset.py ADDED
@@ -0,0 +1,206 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import json
2
+ import os
3
+ import datasets
4
+
5
+
6
+ class COCOBuilderConfig(datasets.BuilderConfig):
7
+
8
+ def __init__(self, name, splits, **kwargs):
9
+ super().__init__(name, **kwargs)
10
+ self.splits = splits
11
+
12
+
13
+ # Add BibTeX citation
14
+ # Find for instance the citation on arxiv or on the dataset repo/website
15
+ _CITATION = """\
16
+ @article{DBLP:journals/corr/LinMBHPRDZ14,
17
+ author = {Tsung{-}Yi Lin and
18
+ Michael Maire and
19
+ Serge J. Belongie and
20
+ Lubomir D. Bourdev and
21
+ Ross B. Girshick and
22
+ James Hays and
23
+ Pietro Perona and
24
+ Deva Ramanan and
25
+ Piotr Doll{'{a} }r and
26
+ C. Lawrence Zitnick},
27
+ title = {Microsoft {COCO:} Common Objects in Context},
28
+ journal = {CoRR},
29
+ volume = {abs/1405.0312},
30
+ year = {2014},
31
+ url = {http://arxiv.org/abs/1405.0312},
32
+ archivePrefix = {arXiv},
33
+ eprint = {1405.0312},
34
+ timestamp = {Mon, 13 Aug 2018 16:48:13 +0200},
35
+ biburl = {https://dblp.org/rec/bib/journals/corr/LinMBHPRDZ14},
36
+ bibsource = {dblp computer science bibliography, https://dblp.org}
37
+ }
38
+ """
39
+
40
+ # Add description of the dataset here
41
+ # You can copy an official description
42
+ _DESCRIPTION = """\
43
+ COCO is a large-scale object detection, segmentation, and captioning dataset.
44
+ """
45
+
46
+ # Add a link to an official homepage for the dataset here
47
+ _HOMEPAGE = "http://cocodataset.org/#home"
48
+
49
+ # Add the licence for the dataset here if you can find it
50
+ _LICENSE = ""
51
+
52
+ # Add link to the official dataset URLs here
53
+ # The HuggingFace dataset library don't host the datasets but only point to the original files
54
+ # This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
55
+
56
+ # This script is supposed to work with local (downloaded) COCO dataset.
57
+ _URLs = {}
58
+
59
+
60
+ # Name of the dataset usually match the script name with CamelCase instead of snake_case
61
+ class COCODataset(datasets.GeneratorBasedBuilder):
62
+ """An example dataset script to work with the local (downloaded) COCO dataset"""
63
+
64
+ VERSION = datasets.Version("0.0.0")
65
+
66
+ BUILDER_CONFIG_CLASS = COCOBuilderConfig
67
+ BUILDER_CONFIGS = [
68
+ COCOBuilderConfig(name='2017', splits=['train', 'valid', 'test']),
69
+ ]
70
+ DEFAULT_CONFIG_NAME = "2017"
71
+
72
+ def _info(self):
73
+ # This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
74
+
75
+ feature_dict = {
76
+ "image_id": datasets.Value("int64"),
77
+ "caption_id": datasets.Value("int64"),
78
+ "caption": datasets.Value("string"),
79
+ "height": datasets.Value("int64"),
80
+ "width": datasets.Value("int64"),
81
+ "file_name": datasets.Value("string"),
82
+ "coco_url": datasets.Value("string"),
83
+ "image_path": datasets.Value("string"),
84
+ }
85
+
86
+ features = datasets.Features(feature_dict)
87
+
88
+ return datasets.DatasetInfo(
89
+ # This is the description that will appear on the datasets page.
90
+ description=_DESCRIPTION,
91
+ # This defines the different columns of the dataset and their types
92
+ features=features, # Here we define them above because they are different between the two configurations
93
+ # If there's a common (input, target) tuple from the features,
94
+ # specify them here. They'll be used if as_supervised=True in
95
+ # builder.as_dataset.
96
+ supervised_keys=None,
97
+ # Homepage of the dataset for documentation
98
+ homepage=_HOMEPAGE,
99
+ # License for the dataset if available
100
+ license=_LICENSE,
101
+ # Citation for the dataset
102
+ citation=_CITATION,
103
+ )
104
+
105
+ def _split_generators(self, dl_manager):
106
+ """Returns SplitGenerators."""
107
+ # This method is tasked with downloading/extracting the data and defining the splits depending on the configuration
108
+ # If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name
109
+
110
+ data_dir = self.config.data_dir
111
+ if not data_dir:
112
+ raise ValueError(
113
+ "This script is supposed to work with local (downloaded) COCO dataset. The argument `data_dir` in `load_dataset()` is required."
114
+ )
115
+
116
+ splits = []
117
+ for split in self.config.splits:
118
+ if split == 'train':
119
+ dataset = datasets.SplitGenerator(
120
+ name=datasets.Split.TRAIN,
121
+ # These kwargs will be passed to _generate_examples
122
+ gen_kwargs={
123
+ "json_path": os.path.join(data_dir, f"captions_train{self.config.name}.json"),
124
+ "image_dir": os.path.join(data_dir, f'train{self.config.name}'),
125
+ "split": "train",
126
+ }
127
+ )
128
+ elif split in ['val', 'valid', 'validation', 'dev']:
129
+ dataset = datasets.SplitGenerator(
130
+ name=datasets.Split.VALIDATION,
131
+ # These kwargs will be passed to _generate_examples
132
+ gen_kwargs={
133
+ "json_path": os.path.join(data_dir, f"captions_val{self.config.name}.json"),
134
+ "image_dir": os.path.join(data_dir, f'val{self.config.name}'),
135
+ "split": "valid",
136
+ },
137
+ )
138
+ elif split == 'test':
139
+ dataset = datasets.SplitGenerator(
140
+ name=datasets.Split.TEST,
141
+ # These kwargs will be passed to _generate_examples
142
+ gen_kwargs={
143
+ "json_path": os.path.join(data_dir, f'image_info_test{self.config.name}.json'),
144
+ "image_dir": os.path.join(data_dir, f'test{self.config.name}'),
145
+ "split": "test",
146
+ },
147
+ )
148
+ else:
149
+ continue
150
+
151
+ splits.append(dataset)
152
+
153
+ return splits
154
+
155
+ def _generate_examples(
156
+ # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
157
+ self, json_path, image_dir, split
158
+ ):
159
+ """ Yields examples as (key, example) tuples. """
160
+ # This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
161
+ # The `key` is here for legacy reason (tfds) and is not important in itself.
162
+
163
+ _features = ["image_id", "caption_id", "caption", "height", "width", "file_name", "coco_url", "image_path", "id"]
164
+ features = list(_features)
165
+
166
+ if split in "valid":
167
+ split = "val"
168
+
169
+ with open(json_path, 'r', encoding='UTF-8') as fp:
170
+ data = json.load(fp)
171
+
172
+ # list of dict
173
+ images = data["images"]
174
+ entries = images
175
+
176
+ # build a dict of image_id -> image info dict
177
+ d = {image["id"]: image for image in images}
178
+
179
+ # list of dict
180
+ if split in ["train", "val"]:
181
+ annotations = data["annotations"]
182
+
183
+ # build a dict of image_id ->
184
+ for annotation in annotations:
185
+ _id = annotation["id"]
186
+ image_info = d[annotation["image_id"]]
187
+ annotation.update(image_info)
188
+ annotation["id"] = _id
189
+
190
+ entries = annotations
191
+
192
+ for id_, entry in enumerate(entries):
193
+
194
+ entry = {k: v for k, v in entry.items() if k in features}
195
+
196
+ if split == "test":
197
+ entry["image_id"] = entry["id"]
198
+ entry["id"] = -1
199
+ entry["caption"] = -1
200
+
201
+ entry["caption_id"] = entry.pop("id")
202
+ entry["image_path"] = os.path.join(image_dir, entry["file_name"])
203
+
204
+ entry = {k: entry[k] for k in _features if k in entry}
205
+
206
+ yield str((entry["image_id"], entry["caption_id"])), entry
download_coco.txt ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ mkdir data
2
+ cd data
3
+ wget http://images.cocodataset.org/zips/train2017.zip
4
+ wget http://images.cocodataset.org/zips/val2017.zip
5
+ wget http://images.cocodataset.org/zips/test2017.zip
6
+ wget http://images.cocodataset.org/annotations/annotations_trainval2017.zip
7
+ wget http://images.cocodataset.org/annotations/image_info_test2017.zip
8
+ unzip train2017.zip
9
+ unzip val2017.zip
10
+ unzip test2017.zip
11
+ unzip annotations_trainval2017.zip
12
+ unzip image_info_test2017.zip
dummy_data/annotations_trainval2017.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:973b53d5b4748ff251816b6d741dcf19c9a94bfa5b38edcf6164592ebd610476
3
+ size 7615
dummy_data/image_info_test2017.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a794a2b342055f4c22d6d7f0a3e29a8b8ec9ffa8b8611b24bd62b49cd84d3ac3
3
+ size 3682
dummy_data/test2017.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:abdc12551acb154ed65c38f493ce930ad22fffd74c1b1e1f811a46212eb28e91
3
+ size 3023535
dummy_data/train2017.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:370a953a7907709727f53c989d8282a1c0e47fb6fc0900cebf1c67ba362c6e73
3
+ size 3590323
dummy_data/val2017.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c3625319b0adc788f918dea4a32a3c551323fd883888e8a03668cb47dd823b94
3
+ size 2556555