Stable Diffusion AI Image Generator

OnlyOneKenobi said:
Some more fun with custom LORA models and Control Net...

View attachment 1484643
Click to expand...
Just trained my first Lora last night using a bunch of sepia photos.

Think i may have trained it too much because after epoch 3 ( out of 15 ) and 40 repeats/image i get a lot of artificing when checking my image grid....

But its a lot of fun to experiment with for sure !.
 
Fjorko said:
Just trained my first Lora last night using a bunch of sepia photos.

Think i may have trained it too much because after epoch 3 ( out of 15 ) and 40 repeats/image i get a lot of artificing when checking my image grid....

But its a lot of fun to experiment with for sure !.
Click to expand...
Yeah, it might be overtrained. I've only trained 1 or 2 LORA "styles" and the rest are faces \ subjects, but generally it works fine for my purposes if I use 1 Epoch for 10-23 or so images.
 
Tried for photo realism for a change, turns out there are a few Dreambooth models that are quite good for that.
This is a custom LORA made from photos of James Hetfield, rendered with the Realistic Vision model and a few control net tricks to get the pose and posture right. It does take quite a bit of inpainting and tweaking to get the likeness close enough to what the person really looks like.

Tallica.jpg
 
Unofficial Photoshop plugin update.

github.com

GitHub - AbdullahAlfaraj/Auto-Photoshop-StableDiffusion-Plugin: A user-friendly plug-in that makes it easy to generate stable diffusion images inside Photoshop using either Automatic or ComfyUI as a backend.

A user-friendly plug-in that makes it easy to generate stable diffusion images inside Photoshop using either Automatic or ComfyUI as a backend. - AbdullahAlfaraj/Auto-Photoshop-StableDiffusion-Plugin
github.com github.com

https://www.reddit.com/r/StableDiffusion/comments/11iuqhv
 
Vorastra said:
Unofficial Photoshop plugin update.

github.com

GitHub - AbdullahAlfaraj/Auto-Photoshop-StableDiffusion-Plugin: A user-friendly plug-in that makes it easy to generate stable diffusion images inside Photoshop using either Automatic or ComfyUI as a backend.

A user-friendly plug-in that makes it easy to generate stable diffusion images inside Photoshop using either Automatic or ComfyUI as a backend. - AbdullahAlfaraj/Auto-Photoshop-StableDiffusion-Plugin
github.com github.com

https://www.reddit.com/r/StableDiffusion/comments/11iuqhv
Click to expand...
The previous version of this plugin wasn't very intuitive, found it easier to just use the Auto1111 GUI. Hopefully it has improved somewhat.
 
For those who like the single line art style, I've trained a LoRA that emulates the style pretty well. Admittedly, not all the images generated using the model are actually true single line drawings, but it still looks pretty cool.

Lineart 03.jpg

You can get it from here : SingleLineArt Style
 
Snipped this pic from the Adobe Firefly article. What's the best way to go about doing something like this at the moment? Feed an image and get color / material iterations of that image. Are the latest IMG2IMG tools suitable?

GT.png
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X