File size: 1,500 Bytes
357fcec
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
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
import datasets

_CITATION = """\
@InProceedings{huggingface:dataset,
title = {Klee},
author={Kevin Li},
year={2022}
}
"""
_DESCRIPTION = """\
128x128 PNG images of kanjis from the Klee font
"""
_HOMEPAGE = "https://huggingface.co/datasets/AlienKevin/klee"
_LICENSE = "CC0"
_REPO = "https://huggingface.co/datasets/AlienKevin/klee"

class ImageSet(datasets.GeneratorBasedBuilder):
    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    'character': datasets.Value("string"),
                    'image': datasets.Image(),
                }
            ),
            supervised_keys=None,
            homepage=_HOMEPAGE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        images_archive = dl_manager.download(f"{_REPO}/resolve/main/images.tgz")
        image_iters = dl_manager.iter_archive(images_archive)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "images": image_iters
                }
            ),
        ]

    def _generate_examples(self, images):
        idx = 0
        for filepath, image in images:
            character = filepath.split('/')[-1][:-4]
            yield idx, {
                "image": {"path": filepath, "bytes": image.read()},
                "character": character,
            }
            idx += 1