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

[Bug]: Batch and Batch Count increment static seeds #3583

Closed
1 task done
a1270 opened this issue Oct 24, 2022 · 4 comments
Closed
1 task done

[Bug]: Batch and Batch Count increment static seeds #3583

a1270 opened this issue Oct 24, 2022 · 4 comments
Labels
bug-report Report of a bug, yet to be confirmed

Comments

@a1270
Copy link

a1270 commented Oct 24, 2022

Is there an existing issue for this?

  • I have searched the existing issues and checked the recent builds/commits

What happened?

When supplied a static generation seed and increasing batch size/count above 1 the following images will be generated with the seed incremented by one.
e.g. using the seed of 42 and batch count of 4 results in images generated with the seeds: 42,43,44,45

Steps to reproduce the problem

  1. Go to txt2img
  2. Set static seed e.g. 42
  3. Set batch or batch count to greater than 1

What should have happened?

All images generated should use provided seed.

Commit where the problem happens

df0a1f8

What platforms do you use to access UI ?

Windows

What browsers do you use to access the UI ?

Mozilla Firefox, Microsoft Edge

Command Line Arguments

--listen --xformers --medvram --deepdanbooru --gradio-img2img-tool color-sketch

Additional information, context and logs

When using static seeds in conjunction with the wildcards addon this is somewhat or a problem. Saying that, the addon itself has an issue with static seeds. AUTOMATIC1111/stable-diffusion-webui-wildcards#2

@a1270 a1270 added the bug-report Report of a bug, yet to be confirmed label Oct 24, 2022
@timntorres
Copy link
Contributor

If it didn't do this, all the outputs would look exactly the same.

@a1270
Copy link
Author

a1270 commented Oct 25, 2022

If it didn't do this, all the outputs would look exactly the same.

Isn't that the whole point of providing a static seed? afaik seeds closer together are not more similar. So it's not like you are getting any benefit by just doing seed+1. Using a static seed with batch now means you are just generating one image based on the input plus batch-1 random.

I specifically want to do this in conjunction with wildcards. I find a good seed and want to churn thru it with other prompts looking for interesting renders. The way i'm doing it now is using x/y plot.

@timntorres
Copy link
Contributor

timntorres commented Oct 25, 2022

I specifically want to do this in conjunction with wildcards. I find a good seed and want to churn thru it with other prompts looking for interesting renders. The way i'm doing it now is using x/y plot.

I see. Historically (since the Stable Diffusion discord bot back in August), Batch/Batch Count has suited the need to generate images from many different seeds given the same prompt. ("Static generation seeds" are useful insofar as they allow you to generate the same assortment of seeds each time.) What you describe seems more appropriate as a feature request than a bug report. Any ideas of how to implement this in a way that preserves current functionality?

@a1270
Copy link
Author

a1270 commented Oct 25, 2022

I see. Historically (since the Stable Diffusion discord bot back in August), Batch/Batch Count has suited the need to generate images from many different seeds given the same prompt. ("Static generation seeds" are useful insofar as they allow you to generate the same assortment of seeds each time.)

So it's a case of mistaken terminology. It's the verb form of 'batch' and not the noun.

What you describe seems more appropriate as a feature request than a bug report. Any ideas of how to implement this in a way that preserves current functionality?

I think it's more of a case of documentation than anything else. Explain that it's a range generation tool and not a duplicator. I see no reason to make code changes as using x/y plot as a duplicator is sufficient. I'll go ahead and close this and hopefully if someone makes the mistake i did they'll stumble upon this.

Thank you for being so helpful and correcting my misunderstanding.

@a1270 a1270 closed this as completed Oct 25, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug-report Report of a bug, yet to be confirmed
Projects
None yet
Development

No branches or pull requests

2 participants