Topaz Photo AI v2.3.1

The raised hand of the girl in red is slightly odd, though.

“black girl playing the blues, roadhouse, crossroads, low light, shallow depth of field, smoky room”

Seed : 848552 | Steps : 25 | Img Width : 832 | Img Height : 512 | Model : photon_v1 | Mode : txt2img | Selected Style : dark fantasy

thanks. By the way, your generation has 6 fingers on one hand

Makes for good pickin’!

Do you mean the original or the edit? I see it on the original.

Note that Leonado AI simply uses the free Stable Difusion templates. Added to this a few models modified for Leonado AI

I also have SD installed locally via Terminal FWIW.

FWIW? What is it?

For what it’s worth


I was looking at a smaller version and didn’t see this. I corrected it, thanks

Just for testing purposes, I first tested your original image to make a light variant. And then improve the hands and eyes

You’re welcome :slight_smile:

Here are some of the primary models that Leonardo uses.
image

dreamshaper v7 is old. Now, here is “dreamshaperXL_sfwTurboDpmppSDE” for Stable Diffusion. And you only need 5 steps.

@plugsnpixels

If I simply use your prompt with my best model, i.e. “dreamshaperXL_sfwTurboDpmppSDE” with your seed, I get this

2 Likes

Good stuff!

See, this is what happens when there is no weekly PAI release to test :wink:

1 Like

Here we specify a guitar in the prompt. Still the same seed


And without retouching

Great, it gave me a motion blur on my hand hahaha

We need to teach AI that guitar strings are always straight!

Nice work though.

Yeah, some things are still a problem. That’s why Loras also exist. They’re models for very specific things. And thanks :wink: