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

Interpretation uncertainty values #69

Open
finnSartoris opened this issue Jan 17, 2022 · 0 comments
Open

Interpretation uncertainty values #69

finnSartoris opened this issue Jan 17, 2022 · 0 comments

Comments

@finnSartoris
Copy link

Thanks for open-sourcing your code!
I have a question about the interpretation of the uncertainty values.

1.) I guess log_var and proj_output should change places like her #34
https://github.com/Halmstad-University/SalsaNext/blob/a02fad97d646d4c132266ab79fbaea3ecfc237ed/train/tasks/semantic/modules/user.py#L154
and proj_argmax = proj_output[0].argmax(dim=0) is missing.

2.) I have had the pretrained model evaluated with the changes listed above and sequence 8 with SalsaNext uncertainty. As output, I do not get any percentage values for epistemic uncertainty. Is it intended that the uncertainty values are not output as percentages, since only the mean variance from the monte carlo sampling is calculated? https://github.com/Halmstad-University/SalsaNext/blob/a02fad97d646d4c132266ab79fbaea3ecfc237ed/train/tasks/semantic/modules/user.py#L159
If so, how can the values be interpreted, except that a higher value reflects a higher uncertainty?
If this was not the intention, I know I have an error somewhere

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