Learning-Deep-Features-for-One-Class-Classification icon indicating copy to clipboard operation
Learning-Deep-Features-for-One-Class-Classification copied to clipboard

Published 20 hours ago • verified publisher icon buiduchanh

How to train my own custom dataset

Open vanchung1995 opened this issue 4 years ago • 1 comments

Hi there, I want to train my custom dataset. But I dont know how dataset structured? Can you give some information? For example, I want to train a model that can detect the image is a dog or not. My data structure like:

data |__dogs |____dog1.png ...... |____dogn.png

I change cp1_normal_path,cp1_test_path,cp1_test_normal,cp1_test_abnormal equals to data/dogs. But the console print: epoch : 1 ,Descriptive loss : nan, Compact loss : nan

I read data.py file but I dont understand cp1_path , cp1_normal_path, cp1_test_path, cp1_test_normal, cp1_test_abnormal means for what?

vanchung1995 avatar Feb 22 '21 10:02 vanchung1995

@vanchung1995 This project seems to need to provide an additional data set as ref data.

Have you found another way to deal with the problem? I have the same problem as you

huangfeng95 avatar Sep 27 '21 08:09 huangfeng95