Detection Training - RuntimeWarning divide by zero #1315
Closed
ArsalanYounus007
started this conversation in
General
Replies: 2 comments 7 replies
-
Is it because the GTs are too close to each other?? I have used the show-sample flag to get this image |
Beta Was this translation helpful? Give feedback.
2 replies
-
One more thing, Even though I let it train with Warnings. I got a checkpoint - I know it might not be a great checkpoint but I thought to load and test it anyways When I load it (using: https://mindee.github.io/doctr/using_doctr/custom_models_training.html) , I get this error. |
Beta Was this translation helpful? Give feedback.
5 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hello,
I am trying to train db_resnet50 for Checkbox (detection). I created my own annotations and started the training.
I am getting
F:\GitHub\doctr\doctr\models\detection\differentiable_binarization\base.py:247: RuntimeWarning: divide by zero encountered in true_divide
Loss goes up whenever this occurs, I am sure that this is a problem with my dataset annotations, But I am unable to figure out what.
Beta Was this translation helpful? Give feedback.
All reactions