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

How do I train on different image dimenstions e.g 192 x 192 instead of 96 x 96 pixels? #55

Open
mornedev opened this issue Jun 29, 2023 · 1 comment

Comments

@mornedev
Copy link

I can train with my own dataset if I use 96 pixels x 96 pixels with matching masks. When I change this to 192 or 512 pixels my model always gets 0.00 IoU. Thanks!

@mornedev
Copy link
Author

Example files:

1
1

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

1 participant