Have a trouble when training model

AI Discussions
,
1

Hello everyone, I’m currently learning about ML and am impressed with computer vision, so I’m doing a small project using CNN to identify whether plants are sick or not. However, when training, I’m having problems.I think It’s overfit, pls help me
colab link:

linked dataset: Kashmiri Apple Plant Disease Dataset | Kaggle

2

I am confused. What does that mean?

4

sorry my bad,it doesn’t make sense

5

The data set is pretty small. And it looks like the model is overfitting the training set.

You already do “dropout.”

Maybe also try

  • augmentation to increase the dataset size
  • stop early. You have 150 epochs. It’s worth trying range(10, 100, 10). For ex. @epoch 100 the accuracy is already > 90%