Visualisasi : Jelajahi di Know Your Data
Deskripsi :
Dataset Mobil berisi 16.185 gambar dari 196 kelas mobil. Data dibagi menjadi 8.144 gambar pelatihan dan 8.041 gambar pengujian, di mana setiap kelas telah dibagi secara kasar dalam pembagian 50-50. Kelas biasanya pada tingkat Make, Model, Tahun, misalnya 2012 Tesla Model S atau 2012 BMW M3 coupe.
Situs web : https://ai.stanford.edu/~jkrause/cars/car_dataset.html
Kode sumber :
tfds.image_classification.Cars196
Versi :
-
2.0.0
: Rilis awal -
2.0.1
: Pembaruan URL situs web -
2.1.0
(default): Memperbaiki bug https://github.com/tensorflow/datasets/issues/3927
-
Ukuran unduhan :
1.82 GiB
Ukuran kumpulan data :
1.85 GiB
Tembolok otomatis ( dokumentasi ): Tidak
Perpecahan :
Membelah | Contoh |
---|---|
'test' | 8.041 |
'train' | 8.144 |
- Struktur fitur :
FeaturesDict({
'bbox': BBoxFeature(shape=(4,), dtype=tf.float32),
'id': Text(shape=(), dtype=tf.string),
'image': Image(shape=(None, None, 3), dtype=tf.uint8),
'label': ClassLabel(shape=(), dtype=tf.int64, num_classes=196),
})
- Dokumentasi fitur :
Fitur | Kelas | Membentuk | tipe D | Keterangan |
---|---|---|---|---|
FiturDict | ||||
kotak | Fitur Kotak | (4,) | tf.float32 | |
Indo | Teks | tf.string | ||
gambar | Gambar | (Tidak ada, Tidak ada, 3) | tf.uint8 | |
label | Label Kelas | tf.int64 |
Kunci yang diawasi (Lihat
as_supervised
doc ):('image', 'label')
Gambar ( tfds.show_examples ):
- Contoh ( tfds.as_dataframe ):
- kutipan :
@inproceedings{KrauseStarkDengFei-Fei_3DRR2013,
title = {3D Object Representations for Fine-Grained Categorization},
booktitle = {4th International IEEE Workshop on 3D Representation and Recognition (3dRR-13)},
year = {2013},
address = {Sydney, Australia},
author = {Jonathan Krause and Michael Stark and Jia Deng and Li Fei-Fei}
}