- Keterangan :
Set data omniglot untuk pembelajaran satu kali. Set data ini berisi 1623 karakter tulisan tangan yang berbeda dari 50 alfabet yang berbeda.
Dokumentasi Tambahan : Jelajahi di Papers With Code
Beranda : https://github.com/brendenlake/omniglot/
Kode sumber :
tfds.image_classification.OmniglotVersi :
-
3.0.0(default): API split baru ( https://tensorflow.org/datasets/splits )
-
Ukuran unduhan :
17.95 MiBUkuran kumpulan data :
12.29 MiBDi-cache otomatis ( dokumentasi ): Ya
Perpecahan :
| Membelah | Contoh |
|---|---|
'small1' | 2.720 |
'small2' | 3.120 |
'test' | 13.180 |
'train' | 19.280 |
- Struktur fitur :
FeaturesDict({
'alphabet': ClassLabel(shape=(), dtype=int64, num_classes=50),
'alphabet_char_id': int64,
'image': Image(shape=(105, 105, 3), dtype=uint8),
'label': ClassLabel(shape=(), dtype=int64, num_classes=1623),
})
- Dokumentasi fitur :
| Fitur | Kelas | Membentuk | Tipe D | Keterangan |
|---|---|---|---|---|
| FiturDict | ||||
| alfabet | Label Kelas | int64 | ||
| alfabet_karakter_id | Tensor | int64 | ||
| gambar | Gambar | (105, 105, 3) | uint8 | |
| label | Label Kelas | int64 |
Kunci yang diawasi (Lihat dokumen
as_supervised):('image', 'label')Gambar ( tfds.show_examples ):

- Contoh ( tfds.as_dataframe ):
- Kutipan :
@article{lake2015human,
title={Human-level concept learning through probabilistic program induction},
author={Lake, Brenden M and Salakhutdinov, Ruslan and Tenenbaum, Joshua B},
journal={Science},
volume={350},
number={6266},
pages={1332--1338},
year={2015},
publisher={American Association for the Advancement of Science}
}