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

Train models to adapt to larger angles #74

Open
qiao89757 opened this issue Oct 18, 2023 · 1 comment
Open

Train models to adapt to larger angles #74

qiao89757 opened this issue Oct 18, 2023 · 1 comment

Comments

@qiao89757
Copy link

Hello, I am very envious of your team for such wonderful work. When I was training the model, I found that the dataset part of this code directly transforms a picture according to the maximum angle of 45 degrees and then inputs it into the model? Then I now need to train a model that can adapt to large angles. I hope you can give me some ideas.Thank you.

@Phil26AT
Copy link
Collaborator

Hi @qiao89757

Thank you! Indeed, the model pretrained on homographies is also more robust to rotations. You could try to add homography adaptation to MegaDepth and see if this translates there.

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