Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Overfitting Detection #22

Open
akirs82 opened this issue May 9, 2023 · 3 comments
Open

Overfitting Detection #22

akirs82 opened this issue May 9, 2023 · 3 comments

Comments

@akirs82
Copy link

akirs82 commented May 9, 2023

Hello Ydwen,

I would like to know, from the training and validation logs how to check for overfitting or underfitting of the model capacity to the dataset ? My training loss is class_dataset type and my validation is pair_dataset type.

regards
akirs

@ydwen
Copy link
Owner

ydwen commented May 9, 2023

Hi akirs82,

You can construct a validation set of pair_dataset type, where the images are from training set.
This is the solution that we are currently using. Will have a push in the near future.

@akirs82 akirs82 closed this as completed May 9, 2023
@akirs82
Copy link
Author

akirs82 commented May 11, 2023

Hello ydwen,
Thank you for the response. Just to summarize. I will split the dataset to training set (as class_dataset) and validation set 1(in the form of pair_dataset type - NON Overlapping with the data in training set). Now i will have validation accuracy and training loss - with this should i get the graphs as below,

  1. Validation1 accuracy vs Iteration
  2. Training loss vs Iteration.
    (Any pointers how these graphs will explain overfitting - usually it is validation accuracy,training accuracy vs iteration or losses vs iteration)

So based on above graphs,
Is this Validation1 accuracy is like a proxy for the training accuracy ?
I also have another dataset which is not part of the training set. Can i use this as well or not for validation set or as a separate validation set 2 ?

regards
akirs

@akirs82 akirs82 reopened this May 11, 2023
@sofpya
Copy link

sofpya commented Jun 19, 2023

Hi, akir82.
I am training my own dataset. how to construct the validation dataset with pair dataset type. Because my original dataset is a classification type dataset.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants