fix: avoid issues when sigma_min is close to 0 #1138
+17
−7
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
daniandtheweb
changed the title
daniandtheweb
changed the title
daniandtheweb
force-pushed
the
fix_kl_optimal
branch
from
da95169 to
f7fa9c6
Compare
daniandtheweb
changed the title
daniandtheweb
force-pushed
the
fix_kl_optimal
branch
from
f7fa9c6 to
f287b5a
Compare
Contributor
Author
|
This still doesn't seem to be 100% reliable. |
Contributor
Author
|
This new change seems to be more reliable. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is a possible fix for #1130.
This PR creates a safeguard that detects whenever a sigma_min gets too close to zero and avoids the issues that may come with it (black images).
This also adds a fix and cleans up the KL Optimal scheduler function.
I've noticed this issue on the Karras and the KL Optimal schedulers with this same model: HassakuXL 3.2.