Datasets:
frgfm
commited on
Commit
·
52861f0
1
Parent(s):
6e52dd9
feat: Added dataset builder
Browse files- imagenette.py +136 -0
imagenette.py
ADDED
@@ -0,0 +1,136 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Copyright (C) 2022, François-Guillaume Fernandez.
|
2 |
+
|
3 |
+
# This program is licensed under the Apache License 2.0.
|
4 |
+
# See LICENSE or go to <https://www.apache.org/licenses/LICENSE-2.0> for full license details.
|
5 |
+
|
6 |
+
"""Imagenette dataset."""
|
7 |
+
|
8 |
+
import os
|
9 |
+
import json
|
10 |
+
|
11 |
+
import datasets
|
12 |
+
|
13 |
+
|
14 |
+
_HOMEPAGE = "https://github.com/fastai/imagenette"
|
15 |
+
|
16 |
+
_LICENSE = "Apache License 2.0"
|
17 |
+
|
18 |
+
_CITATION = """\
|
19 |
+
@software{Howard_Imagenette_2019,
|
20 |
+
title={Imagenette: A smaller subset of 10 easily classified classes from Imagenet},
|
21 |
+
author={Jeremy Howard},
|
22 |
+
year={2019},
|
23 |
+
month={March},
|
24 |
+
publisher = {GitHub},
|
25 |
+
url = {https://github.com/fastai/imagenette}
|
26 |
+
}
|
27 |
+
"""
|
28 |
+
|
29 |
+
_DESCRIPTION = """\
|
30 |
+
Imagenette is a subset of 10 easily classified classes from Imagenet
|
31 |
+
(tench, English springer, cassette player, chain saw, church, French
|
32 |
+
horn, garbage truck, gas pump, golf ball, parachute).
|
33 |
+
"""
|
34 |
+
|
35 |
+
_LABEL_MAP = [
|
36 |
+
'n01440764',
|
37 |
+
'n02102040',
|
38 |
+
'n02979186',
|
39 |
+
'n03000684',
|
40 |
+
'n03028079',
|
41 |
+
'n03394916',
|
42 |
+
'n03417042',
|
43 |
+
'n03425413',
|
44 |
+
'n03445777',
|
45 |
+
'n03888257',
|
46 |
+
]
|
47 |
+
|
48 |
+
|
49 |
+
|
50 |
+
class OpenFireConfig(datasets.BuilderConfig):
|
51 |
+
"""BuilderConfig for OpenFire."""
|
52 |
+
|
53 |
+
def __init__(self, data_url, **kwargs):
|
54 |
+
"""BuilderConfig for OpenFire.
|
55 |
+
Args:
|
56 |
+
data_url: `string`, url to download the zip file from.
|
57 |
+
**kwargs: keyword arguments forwarded to super.
|
58 |
+
"""
|
59 |
+
super(OpenFireConfig, self).__init__(version=datasets.Version("1.0.0"), **kwargs)
|
60 |
+
self.data_url = data_url
|
61 |
+
|
62 |
+
|
63 |
+
class OpenFire(datasets.GeneratorBasedBuilder):
|
64 |
+
"""OpenFire dataset."""
|
65 |
+
|
66 |
+
BUILDER_CONFIGS = [
|
67 |
+
OpenFireConfig(
|
68 |
+
name="full_size",
|
69 |
+
description="All images are in their original size.",
|
70 |
+
data_url="https://s3.amazonaws.com/fast-ai-imageclas/imagenette2.tgz",
|
71 |
+
),
|
72 |
+
OpenFireConfig(
|
73 |
+
name="320px",
|
74 |
+
description="All images were resized on their shortest side to 320 pixels.",
|
75 |
+
data_url="https://s3.amazonaws.com/fast-ai-imageclas/imagenette2-320.tgz",
|
76 |
+
),
|
77 |
+
OpenFireConfig(
|
78 |
+
name="160px",
|
79 |
+
description="All images were resized on their shortest side to 160 pixels.",
|
80 |
+
data_url="https://s3.amazonaws.com/fast-ai-imageclas/imagenette2-160.tgz",
|
81 |
+
),
|
82 |
+
]
|
83 |
+
|
84 |
+
def _info(self):
|
85 |
+
return datasets.DatasetInfo(
|
86 |
+
description=_DESCRIPTION + self.config.description,
|
87 |
+
features=datasets.Features(
|
88 |
+
{
|
89 |
+
"image": datasets.Image(),
|
90 |
+
"label": datasets.ClassLabel(
|
91 |
+
names=[
|
92 |
+
"tench",
|
93 |
+
"English springer",
|
94 |
+
"cassette player",
|
95 |
+
"chain saw",
|
96 |
+
"church",
|
97 |
+
"French horn",
|
98 |
+
"garbage truck",
|
99 |
+
"gas pump",
|
100 |
+
"golf ball",
|
101 |
+
"parachute",
|
102 |
+
]
|
103 |
+
),
|
104 |
+
}
|
105 |
+
),
|
106 |
+
supervised_keys=None,
|
107 |
+
homepage=_HOMEPAGE,
|
108 |
+
license=_LICENSE,
|
109 |
+
citation=_CITATION,
|
110 |
+
)
|
111 |
+
|
112 |
+
def _split_generators(self, dl_manager):
|
113 |
+
data_dir = dl_manager.download_and_extract(self.config.data_url)
|
114 |
+
return [
|
115 |
+
datasets.SplitGenerator(
|
116 |
+
name=datasets.Split.TRAIN,
|
117 |
+
gen_kwargs={
|
118 |
+
"image_folder": os.path.join(data_dir, "train"),
|
119 |
+
"split": "train",
|
120 |
+
},
|
121 |
+
),
|
122 |
+
datasets.SplitGenerator(
|
123 |
+
name=datasets.Split.VALIDATION,
|
124 |
+
gen_kwargs={
|
125 |
+
"image_folder": os.path.join(data_dir, "val"),
|
126 |
+
"split": "validation",
|
127 |
+
},
|
128 |
+
),
|
129 |
+
]
|
130 |
+
|
131 |
+
def _generate_examples(self, image_folder, split):
|
132 |
+
idx = 0
|
133 |
+
for class_idx, class_folder in _LABEL_MAP:
|
134 |
+
for filepath in os.listdir(class_folder):
|
135 |
+
yield idx, {"image": os.path.join(image_folder, class_folder, filepath), "label": class_idx}
|
136 |
+
idx += 1
|