キロニスト

Kuzushiji-MNIST は MNIST データセット (28x28 グレースケール、70,000 画像) のドロップイン置換で、元の MNIST 形式と NumPy 形式で提供されます。 MNIST では 10 クラスに制限されているため、くずし字 MNIST を作成するときに、ひらがなの 10 行のそれぞれを表す 1 文字を選択しました。

スプリット
'test' 10,000
'train' 60,000
  • 機能構造:
FeaturesDict({
    'image': Image(shape=(28, 28, 1), dtype=uint8),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=10),
})
  • 機能のドキュメント:
特徴クラスDtype説明
特徴辞書
画像画像(28, 28, 1) uint8
ラベルクラスラベルint64

視覚化

  • 引用
@online{clanuwat2018deep,
  author       = {Tarin Clanuwat and Mikel Bober-Irizar and Asanobu Kitamoto and Alex Lamb and Kazuaki Yamamoto and David Ha},
  title        = {Deep Learning for Classical Japanese Literature},
  date         = {2018-12-03},
  year         = {2018},
  eprintclass  = {cs.CV},
  eprinttype   = {arXiv},
  eprint       = {cs.CV/1812.01718},
}