10 Stable Diffusion Models Tested With Optimal Settings!

Today I tested 10 different stable diffusion models in an attempt to find the optimal settings including the guidance scale, inference steps, and scheduler that resulted in the "best" image quality.

Here are links to each of the models tested:

[amorgan]

Dude, you save all of us so much time and give us tons of new knowledge. Thanks!

[bussty]

Wow that's so cool, thank you! What's your favourite? Cheers for doing this.

[seanknowles]

I played with the consistent character model and it worked pretty well. Better than most. But I was wondering how to create full body images with consistent character on your platform? My reference image is only a head/shoulders shot and it replicates that, but it wont generate full body images. Is there a way to make that happen?

[theSarinBlack]

Duuude I need an update on your 3d veg tower. I'm building one right now and saw you didn't update much about it last year. Did you retire it? Hope you're well

[masterhand]

first of all great video as always, now plz there is something i cant understand, im using fooocus, and i did the same prompt same everything, but the image is not good as it comes from his site ? can you tell me why plz?

[sander]

When will you make another AI news video? Always looking forward to those ( :

[sergeysaulit]

Can't create an account... Says you've reached your limit

[natieklopper]

I kind of wish these steps could actually be like your previous video and not have an impact when kept the same, somehow the models should have some kind of normalisation state where the steps are scaled from the same constant so to speak.

[STREAM_BOXER]

wont let me create an account. Says you reached your limit