Exactly the same way. Image data is numeric data. When training a vanilla autoencoder (no use of convolutions) on image data, typically the image pixel ...
確定! 回上一頁