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

Smoothing alternative #6830

Open
ivanstepanovftw opened this issue Apr 13, 2024 · 1 comment
Open

Smoothing alternative #6830

ivanstepanovftw opened this issue Apr 13, 2024 · 1 comment

Comments

@ivanstepanovftw
Copy link

I want to see alternative to smoothing which aligns to minimum/maximum observed value at this step. This method would help in better capturing the extreme values that might be critical for fine-tuning and diagnosing the training process.

For example, no smoothing and suggested feature:
before

after

@arcra
Copy link
Member

arcra commented Apr 18, 2024

Hi Ivan, that seems like a reasonable feature, tho I don't know how often it would be used, and how much it would clutter our settings panel, which is already reasonably dense. I'll bring it up to the team and figure if we'd like to add something like this and if we have the bandwidth to do this.

Would you be interested in contributing to our code base? Then I can check if we have the bandwidth to review the request and to review the code changes.

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