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

About Val #11

Open
celiazzx opened this issue Jul 28, 2022 · 3 comments
Open

About Val #11

celiazzx opened this issue Jul 28, 2022 · 3 comments

Comments

@celiazzx
Copy link

Hi, did you do the verification. If so how was the validation set organized

@daooshee
Copy link
Owner

Hi, you can refer to this issue.

@celiazzx
Copy link
Author

It seems that your training process does not do validation. When I run val_ss.py I get an error: 'Error in WIDERDetectionTransfer: XXX not found', I think I should also need a SRC_VAL_FILE, which corresponds to wider original and wider add noise respectively

@daooshee
Copy link
Owner

daooshee commented Aug 1, 2022

It seems that your training process does not do validation.

Yes, I split training and validation for convenience. I first save many checkpoints (each 2500 iteration) and then do validation to find the best one.

When I run val_ss.py I get an error: 'Error in WIDERDetectionTransfer: XXX not found', I think I should also need a SRC_VAL_FILE, which corresponds to wider original and wider add noise respectively

It indeed needs WIDER FACE validation set with noise. I will try to fix this problem and upload the WIDER FACE validation set as soon as possible. You can also create this set using released models following ./train_noise_synthesis or use SRC_TRAIN_FILE to replace SRC_VAL_FILE in val_loader_ss_consrc.

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

2 participants