binarized_mnist

  • 説明

最初に使用された MNIST 画像の特定の 2 値化 (Salakhutdinov & Murray、2008)。このデータセットは画像の生成モデルを評価するために頻繁に使用されるため、ラベルは提供されていません。

スプリット
'test' 10,000
'train' 50,000
'validation' 10,000
  • 機能の構造:
FeaturesDict({
   
'image': Image(shape=(28, 28, 1), dtype=uint8),
})
  • 機能ドキュメント:
特徴クラスDタイプ説明
特徴辞書
画像画像(28、28、1) uint8

視覚化

  • 引用
@inproceedings{salakhutdinov2008quantitative,
title
={On the quantitative analysis of deep belief networks},
author
={Salakhutdinov, Ruslan and Murray, Iain},
booktitle
={Proceedings of the 25th international conference on Machine learning},
pages
={872--879},
year
={2008},
organization
={ACM}
}