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

how to use random seed for sample? #350

Closed
tcflying opened this issue Mar 10, 2023 · 3 comments
Closed

how to use random seed for sample? #350

tcflying opened this issue Mar 10, 2023 · 3 comments

Comments

@tcflying
Copy link

i got error when i use prompt like:

a beautiful girl -w 640 -h 1280 -7 -s 16 -d -1

image

then will generate 512*512 pictures.

@DKnight54
Copy link
Contributor

Not sure if it will fix the 512 by 512 issue, but from a quick scan of Kohya's code, it will default to random seed if there is no -d arguement. In other words? Instead of "a beautiful girl -w 640 -h 1280 -7 -s 16 -d -1", feed the sample generator "a beautiful girl -w 640 -h 1280 -7 -s 16". That's what I do.

As for the error message, I'm not sure it's related to your issue with the image size being stuck at 512x512, but it's saying that the version of DDIM scheduler that Kohya's including in the SD-scripts has incorrect configs. Since AFAIK, they are having the script automatically pull it from Hugging Face, either Hugging Face needs to do some changes, or maybe Kohya needs to amend their code to check if the checksums match for the current version, and pull new ones if there are any?

@tcflying tcflying reopened this Mar 13, 2023
@tcflying
Copy link
Author

thanks for the help, i removed -d, but the sample picture were still 512*512, if you working fine?

@bmaltais
Copy link
Owner

Use --w and --h to set the width and height

Cauldrath pushed a commit to Cauldrath/kohya_ss that referenced this issue Apr 5, 2023
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

3 participants