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

Followup for multi-annotation trainings #8097

Open
1 of 4 tasks
daniel-wer opened this issue Sep 24, 2024 · 0 comments
Open
1 of 4 tasks

Followup for multi-annotation trainings #8097

daniel-wer opened this issue Sep 24, 2024 · 0 comments
Assignees

Comments

@daniel-wer
Copy link
Member

daniel-wer commented Sep 24, 2024

Detailed Description

#8071 allows starting a job to train a model on multiple annotations, selected by id or uri. Follow-up wishes came up:

  • When training a model from within wk, currently all bounding boxes of all annotations need to have the same size. This restricts which training data can be combined. Slight variations in size are usually no big deal and one even sometimes decides to "waste" some training data, because that's better than not using the data at all. Also, this becomes even more relevant with the next point. This will also require changes in the worker which should use the smallest bounding box of all to guide the training_sample_size computation (and related values).
  • It should be possible to select a resolution for each annotation which is the resolution that will be trained on. Many a time this is not the finest resolution (which is chosen currently). This is a frontend only issue.

Context

  • Specific to long-running jobs (set jobsEnabled=true in application.conf)
  • Specific to webknossos.org (set isDemoInstance=true in application.conf)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants