Construct CompVisDenoiser with quantization for better output #460
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
|
This doesn't feel overall like an upgrade plus it changes pictures for existing seeds. Enabled Disabled |
|
I disagree - I think it's better than the default - if you look at your linked images, for example, row 3 column 2, quantization has produced a much more coherent hand. It's also universally sharper and removes quite a bit of the noise compared to default. I could make it a cmdline option? |
|
Maybe make it an option similar to the color correction option in the settings menu. |
|
yeah, add it to settings |
|
@AUTOMATIC1111 I've added a setting for this. |
Sashimimochi
pushed a commit
to Sashimimochi/stable-diffusion-webui
that referenced
this pull request
Sep 22, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 PR enables quantization in CompVisDenoiser for K samplers for better/cleaner/sharper output. See Birch-san/stable-diffusion@f52429a and crowsonkb/k-diffusion#23
Comparisons (zoomed %400)
Before:
After:
