- Описание :
Набор данных с изображениями из 5 классов (информацию о конкретном классе см. В названии конфигурации)
Домашняя страница : https://www.kaggle.com/c/petfinder-adoption-prediction/data
Исходный код :
tfds.image_classification.PetFinder
Версии :
-
1.0.0
(по умолчанию): без примечаний к выпуску.
-
Размер загружаемого файла :
1.94 GiB
Размер набора данных :
Unknown size
Автоматическое кэширование ( документация ): неизвестно
Сплит :
Трещина | Примеры |
---|---|
'test' | 14 465 |
'train' | 58 311 |
- Особенности :
FeaturesDict({
'PetID': Text(shape=(), dtype=tf.string),
'attributes': FeaturesDict({
'Age': tf.int64,
'Breed1': tf.int64,
'Breed2': tf.int64,
'Color1': tf.int64,
'Color2': tf.int64,
'Color3': tf.int64,
'Dewormed': tf.int64,
'Fee': tf.int64,
'FurLength': tf.int64,
'Gender': tf.int64,
'Health': tf.int64,
'MaturitySize': tf.int64,
'Quantity': tf.int64,
'State': tf.int64,
'Sterilized': tf.int64,
'Type': tf.int64,
'Vaccinated': tf.int64,
'VideoAmt': tf.int64,
}),
'image': Image(shape=(None, None, 3), dtype=tf.uint8),
'image/filename': Text(shape=(), dtype=tf.string),
'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=5),
})
Контролируемые ключи (см.
as_supervised
doc ):('attributes', 'label')
Цитата :
@ONLINE {kaggle-petfinder-adoption-prediction,
author = "Kaggle and PetFinder.my",
title = "PetFinder.my Adoption Prediction",
month = "april",
year = "2019",
url = "https://www.kaggle.com/c/petfinder-adoption-prediction/data/"
}
- Рисунок ( tfds.show_examples ):
- Примеры ( tfds.as_dataframe ):