[go: nahoru, domu]
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.

Sign up for GitHub

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

Jump to bottom

a naive question about crf. #15

Closed
ghost opened this issue Jun 25, 2018 · 1 comment
Closed

a naive question about crf. #15

ghost opened this issue Jun 25, 2018 · 1 comment

Comments

@ghost
Copy link
ghost commented Jun 25, 2018

Thanks for your excellent code!
I am a novice at crf. I notice that crf is used during evaling but not during training. Could someone tell me why? Thanks for your answer!
@kazuto1011
Copy link
Owner
kazuto1011 commented Jun 25, 2018 •
edited
Loading

I think the FCN training is like a part of CRF optimization. The CRF post-processing in evaluation is actually decoupled optimization using log probabilities from the fixed FCN (mean field inference with a fixed unary potential). I followed the DeepLab v2 way. See section 3.3 of the original paper.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@kazuto1011