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

Random strength for Noise Offset and input perturbation noise #1177

Merged
merged 3 commits into from
Mar 20, 2024

Conversation

KohakuBlueleaf
Copy link
Contributor

As title, I implemented a mechanism to apply noise offset/ip noise with random strength. (based on the strength value)

This is simple and could help model to learn more stably

@masaint
Copy link

masaint commented Mar 18, 2024

I'm interested in this concept, however is the random value truly 0-1? If so, have you tested it? Would be interested to see results.

Perhaps it could be improved if allowing a random range rather than truly random?

@KohakuBlueleaf
Copy link
Contributor Author

I'm interested in this concept, however is the random value truly 0-1? If so, have you tested it? Would be interested to see results.

Perhaps it could be improved if allowing a random range rather than truly random?

image

And have you checked the code? it is random range. the upper bound is the strength set in the arguments.

@kohya-ss
Copy link
Owner

Thank you for this simple but really effective PR!

@kohya-ss kohya-ss merged commit cf09c6a into kohya-ss:dev Mar 20, 2024
1 check passed
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

Successfully merging this pull request may close these issues.

3 participants