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

I cannot reproduce vit-small@cocostuff27 using the default config #63

Open
rayleizhu opened this issue Dec 14, 2022 · 1 comment
Open

Comments

@rayleizhu
Copy link

Hi, thanks for open-sourcing this elegant codebase.

As a starting point, I try to reproduce the results in your paper. However, using vit-small@cocostuff27, I can only achieve 18.13/30.59 mIoU (no CRF) using cluster/linear probe respectively, while the reported results in the paper are 24.0/38.4 mIoU (no CRF). I do not change anything in the code, and only change the 6 hyperparameters to the numbers reported in the paper.

How can I reproduce the performance mentioned in the paper?

@mhamilton723
Copy link
Owner

If you load our pretrained model you should able to reproduce those numbers. The pre-trained model also has the specific values of the hyperparameters used so that you can verify. Also the CRF adds a few points of perf in there too

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