I am trying to classify normal faces and faces with snow app that put a variety
of stickers on.
Here is code snippets and result screenshots.
The above results look great to me. Validation loss is lower than training loss
and the accuracy is high. But, classification seems not to work correctly.
Here are two folders of image dataset.
How can I fix this result correctly? I will appreciate it if you could give me a
n advise to address this.