[Stable Diffusion] Prompt Sharing and Learning Thread

[sharlotte]

I'm not sure whether the restore face is there any longer. A very good tool for face is adetailer in Automatic1111. Just look for adetailer in the 'extensions' tab to install. There are loads of videos about it - this one is for the install for instance (

You must be registered to see the links

).

 

[namhoang909]

I'm not sure whether the restore face is there any longer. A very good tool for face is adetailer in Automatic1111. Just look for adetailer in the 'extensions' tab to install. There are loads of videos about it - this one is for the install for instance (

You must be registered to see the links

).

I think I've found the face restore, long story: I tried XYZ plot and saw "Face Restore" so I chose it and input some random value to see if it worked, checking the cmd I saw SD was downloading from this link

You must be registered to see the links

, not sure if it can be used in Automatic1111 though

 

[Jimwalrus]

"

You must be registered to see the links

"

I think I've found the face restore, long story: I tried XYZ plot and saw "Face Restore" so I chose it and input some random value to see if it worked, checking the cmd I saw SD was downloading from this link

You must be registered to see the links

, not sure if it can be used in Automatic1111 though

For Face Restore, I'd recommend using GFPGAN over CodeFormers - far better, especially for photorealistic. In A1111, select Face Restoration on the Settings tab. I like to use it at ~0.05 visibility.

 

[Sepheyer]

I think I've found the face restore, long story: I tried XYZ plot and saw "Face Restore" so I chose it and input some random value to see if it worked, checking the cmd I saw SD was downloading from this link

You must be registered to see the links

, not sure if it can be used in Automatic1111 though

And if I may with my $0.02 - do switch to ComfyUI rather than A1111. I know at first it will be butthurt for weeks on end, but it will pay off.

 

[Jimwalrus]

And if I may with my $0.02 - do switch to ComfyUI rather than A1111. I know at first it will be butthurt for weeks on end, but it will pay off.

If I wanted butthurt for weeks on end, I'd forget my wife's birthday again...

 

[SDAI-futa]

There are extensions that let you build entire prompt libraries with image previews.

Can you expand on that a bit? What extension that you know help with building images from img2img? I use Pose Net a lot (with different models) and it helps for sure, but always happy to learn of more. Also techniques. Thanks!

 

[DreamingAway]

Can you expand on that a bit? What extension that you know help with building images from img2img? I use Pose Net a lot (with different models) and it helps for sure, but always happy to learn of more. Also techniques. Thanks!

In that quote I meant to say that you can save images with meta data into a local library that is indexed and searchable. It's just a much better user experience than the built in A111 prompt library (which is extremely old.. that prompt tool was in the very first version of A111 and predated all the amazing stuff we have today)

 

[Synalon]

This is my current ui, a couple questions
where is the "Restore face" setting?
while I have indicate in both positive + negative it still generated a nude picture, & sometimes other people appear, how to fix that? thank in advance. View attachment 3188852

Spoiler: TI + Lora

You don't have permission to view the spoiler content. Log in or register now.

Spoiler: Lora only

You don't have permission to view the spoiler content. Log in or register now.

Spoiler: TI only

You don't have permission to view the spoiler content. Log in or register now.

You can find restore face by going to settings.

Settings - face restoration (in post processing) - tick the box

As others have said there are extensions for codeformer/gfpgan to make it show in the UI.

 

[SDAI-futa]

There are extensions that let you build entire prompt libraries with image previews.

In that quote I meant to say that you can save images with meta data into a local library that is indexed and searchable. It's just a much better user experience than the built in A111 prompt library (which is extremely old.. that prompt tool was in the very first version of A111 and predated all the amazing stuff we have today)

Actually, share that to then please. Would help.

 

[DreamingAway]

Actually, share that to then please. Would help.

There are many.

Example:

You must be registered to see the links

 

[Mr-Fox]

I would recommend to only use face restore if absolutely necessary. It used to be needed but since all checkpoints and SD in general has improved most of the time you get better result without face restore. However you can use postprocessing instead and again GFPGAN is superior to codeformer in most cases. You can find it in settings/postprocessing. Select both GFPGAN and Codeformer if you wish, this will add an option in txt2img and img2img tab for postprocessing. They have sliders so you can finetune them and even blend them if you wish, though GFPGAN is way better on it's own over codeformer. I use GFPGAN postprocessing in all my images. It will improve the face and in particular the eyes. If you combine it with after detailer "mediapipe_face_mesh_eyes_only" you can get really stunning eyes.

No face restore and no postprocessing	GFPGAN postprocessing	GFPGAN postprocessing and adetailer

I didn't manage to demonstrate the effect of adetailer as well as I wanted but take my word for it that it is very useful, though this checkpoint (devlishphotorealism) does produce very nice eyes. With lesser models the effect might be more obvious.

Now that you have a nice face you can use this with either roop or reactor for faceswapping or ip adapter in controlnet for a more consistent character.

To avoid getting other people in your image use tags such as "solo" and in negative "bystanders" or similair. The image ratio and resolution might produce twins and how the model have been trained can produce other people in the image as well. Using embeddings or loras such as easy negative or similar have a tendency to give unwanted side effects so this could also be the cause.

In regards to nudes or dressed this has also to do with how the model have been trained. You can try "NSFW" in negative and/or "undressed" and in positive try "dressed". Instead of using "big round breasts" use "big round bust" as this indicates dressed breasts.

 

[Mr-Fox]

I forgot to mention that the image ratio format can have an impact on the general image quality. Most often it gives better result using portrait format image rather than a square 1:1 image. I recommend to use either 512x786 or 640x960. It's better if your pc can run the larger resolution. Clipskip 2 is very commonly used and also in training the models, it can have a profound effect on the image quality and also fix things such as multiple navals etc, highly recommended though not always.


This image demonstrates adetailer better, I also used clipskip 2.



I made a similar image to namhoang909´s prompt and in the 2nd I used the face portrait with reactor for the faceswap. If it's not to your taste this is only a demo.

 

[Mr-Fox]

There has been people asking about creating consistent characters and changing outfit etc.
I used the image in my previous post in open pose and changed the outfit prompt and again the face in reactor.
This is one solution, there are also others of course.

 

[picobyte]

My eyes are up here, pervert! challenge:

If you won you may open:

Spoiler: more images

You don't have permission to view the spoiler content. Log in or register now.

 

[Mr-Fox]

Love this one.

 

[picobyte]

Love this one.

I posted some more from the same batch

You must be registered to see the links

, and a

You must be registered to see the links

one, (below images, others on civitai):

 

[sharlotte]

My eyes are up here, pervert! challenge:
View attachment 3191159 View attachment 3191172 View attachment 3191174 View attachment 3191177 View attachment 3191178 View attachment 3191179 View attachment 3191180

I loved all those pussies: very cute and fluffy!

 

[Sharinel]

I posted some more from the same batch

You must be registered to see the links

, and a

You must be registered to see the links

one, (below images, others on civitai):
View attachment 3191917 View attachment 3191918 View attachment 3191920 View attachment 3193037 View attachment 3193039 View attachment 3193046

I liked the look of these and tweaked them with my some of the styles I have saved. Some nice outcomes, posted png's so hopefully folks can see just how they differed from Picos

 

[picobyte]

Well done

 

[modine2021]

dont know what's the deal. but all my images coming out like this. no matter the model. even with default settings and simple prompts