[Stable Diffusion] Prompt Sharing and Learning Thread

[Jimwalrus]

I'm so glad I stumbled upon this thread. I just started playing around with Stable Diffusion a couple of days ago and have kind of got addicted to trying to get beautiful images from it. I'm using img2img and have been experimenting with things like bokeh etc.

My biggest problem isn't the extra limbs/duplicate heads anymore since I filled out the negative keywords more thoroughly and kept size to 512 x 512, it's the eyes. Out of 20 images, only a couple will have normalish eyes, the rest are pretty much mutant or like runny eggs

I started using inpaint and doing a mask over the eyes which has helped a bit but not completely. The link for the beginners guide stuff is great, I'm going to read through all that.

Have I missed something or is there a way to save all your settings in the Web UI without having to fill the keywords etc. in every time I launch it?

Are you using Hires.fix? If not, it's best to switch that on, even if you only upscale a small amount (~10% or so). It basically gives SD another go at generating, honing the result further. For photorealistic use a denoising strength between 0.25 and 0.33. Cartoon/anime can go a little higher.
Use ~1.5x the number of Sampling Steps i.e. if you have 30 sampling steps, use 45 Hires steps.

 

[Mr-Fox]

I'm so glad I stumbled upon this thread. I just started playing around with Stable Diffusion a couple of days ago and have kind of got addicted to trying to get beautiful images from it. I'm using img2img and have been experimenting with things like bokeh etc.

My biggest problem isn't the extra limbs/duplicate heads anymore since I filled out the negative keywords more thoroughly and kept size to 512 x 512, it's the eyes. Out of 20 images, only a couple will have normalish eyes, the rest are pretty much mutant or like runny eggs

I started using inpaint and doing a mask over the eyes which has helped a bit but not completely. The link for the beginners guide stuff is great, I'm going to read through all that.

Have I missed something or is there a way to save all your settings in the Web UI without having to fill the keywords etc. in every time I launch it?

It's easier for us to help if you post an image so we can see what you are working with. In the output folders grab the png file and upload it here. We will load it into png Info and can then see all the settings.
Can you explain more, why only img2img?
If your GPU can handle it, use at least 512x768. SD does better in general with portrait ratio, though it can do lanscape but it's not as easy. I would recommend 640x960. The higher the resolution from the start the more detail you can get in the end result regardless of any steps you take after the initial generation.
A tip is to use postprocessing GFPGAN. This helps with the eyes and the face. Don't confuse it with face restoration though.
The next thing is to use after detailer with the "mediapipe_face_mesh_eyes_only" model. I typically use inpaint denoising strength 0.22 .

You can save prompts as "style" here:


Or you can load the last prompt and settings with the diagonal white arrow with blue background, or do as I described above. Load a png file with png Info and send to txt2img or img2img.

You must be registered to see the links

 

[Lun@]

Are you using Hires.fix? If not, it's best to switch that on, even if you only upscale a small amount (~10% or so). It basically gives SD another go at generating, honing the result further. For photorealistic use a denoising strength between 0.25 and 0.33. Cartoon/anime can go a little higher.
Use ~1.5x the number of Sampling Steps i.e. if you have 30 sampling steps, use 45 Hires steps.

Oh, I wasn't using Hires.fix, will try that! my denoiser strength was too high as well, I had it at 0.7
Sampling was on at 40.

Thanks for the tips! it's quite addictive seeing how good you can get the images

 

[Lun@]

It's easier for us to help if you post an image so we can see what you are working with. In the output folders grab the png file and upload it here. We will load it into png Info and can then see all the settings.
Can you explain more, why only img2img?
If your GPU can handle it, use at least 512x768. SD does better in general with portrait ratio, though it can do lanscape but it's not as easy. I would recommend 640x960. The higher the resolution from the start the more detail you can get in the end result regardless of any steps you take after the initial generation.
A tip is to use postprocessing GFPGAN. This helps with the eyes and the face. Don't confuse it with face restoration though.
The next thing is to use after detailer with the "mediapipe_face_mesh_eyes_only" model. I typically use inpaint denoising strength 0.22 .

You can save prompts as "style" here:

View attachment 3293829
Or you can load the last prompt and settings with the diagonal white arrow with blue background, or do as I described above. Load a png file with png Info and send to txt2img or img2img.

You must be registered to see the links

This was one of my images from the output folder:



I was using img2img because I wanted to generate a whole bunch of variations from a single image. I'm really new to this stuff, so I'm really at the basic level

My gfx card is an rtx 3090 so it should be fine.

Ah thats how I save it, thanks!

 

[Sepheyer]

I'm so glad I stumbled upon this thread. I just started playing around with Stable Diffusion a couple of days ago and have kind of got addicted to trying to get beautiful images from it. I'm using img2img and have been experimenting with things like bokeh etc.

My biggest problem isn't the extra limbs/duplicate heads anymore since I filled out the negative keywords more thoroughly and kept size to 512 x 512, it's the eyes. Out of 20 images, only a couple will have normalish eyes, the rest are pretty much mutant or like runny eggs

I started using inpaint and doing a mask over the eyes which has helped a bit but not completely. The link for the beginners guide stuff is great, I'm going to read through all that.

Have I missed something or is there a way to save all your settings in the Web UI without having to fill the keywords etc. in every time I launch it?

For image to image the size of the canvas matters a lot. If you start with a 512x512 with a fullshot image you want to upscale the latent (note, not the image, but the latent) a few times (say once by 1.5 and then one more time by 1.5 bringing it to 1152) so you can get the face rendered correctly.

In a sense this is a well known problem (i.e. anyone who tried doing what you are doing ran into it, including myself). And the least frustrating approach to resolving it is to switch to ComfyUI so you can clearly see what your workflow is and then use one of the i2i workflows that are posted througout this thread. I think the bulk of these workflows is around September of 2023 when the ComfyUI went mainstream.

Here's an illustration of how i2i takes the small image and scales it up:

Spoiler: CUI Prompt

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

 

[Jimwalrus]

For image to image the size of the canvas matters a lot. If you start with a 512x512 with a fullshot image you want to upscale the latent (note, not the image, but the latent) a few times (say once by 1.5 and then one more time by 1.5 bringing it to 1152) so you can get the face rendered correctly.

In a sense this is a well known problem (i.e. anyone who tried doing what you are doing ran into it, including myself). And the least frustrating approach to resolving it is to switch to ComfyUI so you can clearly see what your workflow is and then use one of the i2i workflows that are posted througout this thread. I think the bulks of these workflows are around September of 2022 when the ComfyUI went mainstream.

Here's an illustration of how i2i takes the small image and scales it up:
View attachment 3293874

Spoiler: CUI Prompt

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

"Why not try ComfyUI?"
"ComfyUI can fix that"
"Go on, try ComfyUI..."

Don't switch! You'll go mad and all your dreams will be of spaghetti and string!

 

[Lun@]

For image to image the size of the canvas matters a lot. If you start with a 512x512 with a fullshot image you want to upscale the latent (note, not the image, but the latent) a few times (say once by 1.5 and then one more time by 1.5 bringing it to 1152) so you can get the face rendered correctly.

In a sense this is a well known problem (i.e. anyone who tried doing what you are doing ran into it, including myself). And the least frustrating approach to resolving it is to switch to ComfyUI so you can clearly see what your workflow is and then use one of the i2i workflows that are posted througout this thread. I think the bulks of these workflows are around September of 2022 when the ComfyUI went mainstream.

Here's an illustration of how i2i takes the small image and scales it up:
View attachment 3293874

Spoiler: CUI Prompt

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

Wow, a lot of great tips from you guys and I have a good bit of reading up still to do it seems

I'm going to compile these tips into a document for my reference while I try these suggestions.

Thanks for this info!

 

[Lun@]

"Why not try ComfyUI?"
"ComfyUI can fix that"
"Go on, try ComfyUI..."

Don't switch! You'll go mad and all your dreams will be of spaghetti and string!

I'm just trying to get to grips with one Ui as it is..

I have to go play with my font of new found knowledge before my dreams become spaghetti and string

 

[Mr-Fox]

For image to image the size of the canvas matters a lot. If you start with a 512x512 with a fullshot image you want to upscale the latent (note, not the image, but the latent) a few times (say once by 1.5 and then one more time by 1.5 bringing it to 1152) so you can get the face rendered correctly.

In a sense this is a well known problem (i.e. anyone who tried doing what you are doing ran into it, including myself). And the least frustrating approach to resolving it is to switch to ComfyUI so you can clearly see what your workflow is and then use one of the i2i workflows that are posted througout this thread. I think the bulk of these workflows is around September of 2023 when the ComfyUI went mainstream.

Here's an illustration of how i2i takes the small image and scales it up:
View attachment 3293874

Spoiler: CUI Prompt

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

Have you heard of the good news? Your lord and savior spagetti the almighty have come to earth. Now you can generate clear too.. And save your Thetan. Just add a this and that node and a handfull of the other plugin, just watch out for those pesky suppressives..

 

[Mr-Fox]

This was one of my images from the output folder:

View attachment 3293865

I was using img2img because I wanted to generate a whole bunch of variations from a single image. I'm really new to this stuff, so I'm really at the basic level

My gfx card is an rtx 3090 so it should be fine.

Ah thats how I save it, thanks!

You can use controlnet with openpose and simply switch checkpoint model and change a few prompt tags for different backdrops and style etc. There are many ways to generate variations on the same image.
I love this "look". Very "gothic" or "black metal". I see what you mean about the eyes.

 

[Lun@]

You can use controlnet with openpose and simply switch checkpoint model and change a few prompt tags for different backdrops and style etc. There are many ways to generate variations on the same image.
I love this "look". Very "gothic" or "black metal". I see what you mean about the eyes.

Yeah I was doing batch files of 20 and getting maybe 1 or 2 decentish ones with my limited knowledge.

Thanks! this was my favourite of them all:



Far from perfect but the overall look was nice

 

[Jimwalrus]

I'm just trying to get to grips with one Ui as it is..

I have to go play with my font of new found knowledge before my dreams become spaghetti and string

In all seriousness, I'm sure even Sepheyer would recommend getting to grips with SD in A1111 first, then transitioning that knowledge to CUI.

Just as you'd learn to fly in a Cessna, then progress to a Gulfstream.
The power and control of CUI looks so tempting, but the idea of facing a 2nd learning curve when all I want is to produce hotties who look like celebs (or just hot) puts me off.

 

[Lun@]

In all seriousness, I'm sure even Sepheyer would recommend getting to grips with SD in A1111 first, then transitioning that knowledge to CUI.

Just as you'd learn to fly in a Cessna, then progress to a Gulfstream.
The power and control of CUI looks so tempting, but the idea of facing a 2nd learning curve when all I want is to produce hotties who look like celebs (or just hot) puts me off.

I feel I have so much to learn about SD at the moment that anything else will be on the backburner for now.
I've literally spent only a day using SD at this point and it was mostly experimenting

 

[namhoang909]

"Why not try ComfyUI?"
"ComfyUI can fix that"
"Go on, try ComfyUI..."

Don't switch! You'll go mad and all your dreams will be of spaghetti and string!

it is quite unfortunate that I have not successfully generated any acceptable image in CUI(even tried efficient pack that gives similar weight as A1111), so while I am interesting in it, upscaling and experimenting are the only things I have done so far.

 

[me3]

In all seriousness, I'm sure even Sepheyer would recommend getting to grips with SD in A1111 first, then transitioning that knowledge to CUI.

Just as you'd learn to fly in a Cessna, then progress to a Gulfstream.
The power and control of CUI looks so tempting, but the idea of facing a 2nd learning curve when all I want is to produce hotties who look like celebs (or just hot) puts me off.

For me there's two main reasons why i use comfy much more than a1111, primarily it's due to the differences in memory usage, i haven't checked in 1.7 but when i can't even fully load an XL model and using controlnet (and other such addons that affect memory usage) without limiting image size for SD15, it gets easier to "use something else".
Second reasons is i generally like to "put stuff together" and experiment to see how things fit/work, which in comfy is much more of a moving puzzle pieces around than having to write code for a1111.

a1111 is a perfectly fine tool for the job, with its faults, same as everything else, and i'd happily use it more if i could. If you don't have the same vram concerns as me (or worse) the choice is probably a bit less straight forward and it's more "i like this one and it does what i need". Free will and options is good

 

[Mr-Fox]

A few examples of creating variations of essentially the same image in txt2img with openpose. Img2img can be vary useful but for this scenario txt2img is much more capable imo. I leave most settings and prompt (composition, character and pose) the same and only change the context, weather, scenery and outfit. I also switch checkpoint model and upscaler for hiresfix.

I "borrowed" the trenchcoat lady by the eminent Thalies. But since the prompt wasn't included I whipped up a prompt myself.



The prompt and data is included in each png file. Just load it in png info, then send it to txt2img etc.



A few tips:

- Keep the prompt simple and don't use the shotgun approach by adding a ton of tags and phrases without reason.
Don't add too many loras and/or Ti's. Be methodical and only add one element at the time otherwise you don't know what effects what.

- Don't copy the prompt practices you find on civitai etc.
People there doesn't know what they are doing most of the time and you see a lot of shot gunning and throwing everything against the wall to see what sticks.
The images there are of course hand picked and is not at all representative of their workflow or process.

- Don't keep banging your head against the wall.
I very rarely do batches. If you don't get the result you want within a few tries with prompt adjustments etc, then switch checkpoint model. Don't try to force it to do something it has not been trained to do.
The same goes for issues with the eyes etc. Either try a different ckpt or simply fix it after with inpainting etc.
You could of course use an extention like after detailer etc to give SD a helping hand.

 

[Thalies]

A few examples of creating variations of essentially the same image in txt2img with openpose. Img2img can be vary useful but for this scenario txt2img is much more capable imo. I leave most settings and prompt (composition, character and pose) the same and only change the context, weather, scenery and outfit. I also switch checkpoint model and upscaler for hiresfix.

I "borrowed" the trenchcoat lady by the eminent Thalies. But since the prompt wasn't included I whipped up a prompt myself.

View attachment 3294318 View attachment 3294405

The prompt and data is included in each png file. Just load it in png info, then send it to txt2img etc.

View attachment 3294344 View attachment 3294343 View attachment 3294345 View attachment 3294352

A few tips:

- Keep the prompt simple and don't use the shotgun approach by adding a ton of tags and phrases without reason.
Don't add too many loras and/or Ti's. Be methodical and only add one element at the time otherwise you don't know what effects what.

- Don't copy the prompt practices you find on civitai etc.
People there doesn't know what they are doing most of the time and you see a lot of shot gunning and throwing everything against the wall to see what sticks.
The images there are of course hand picked and is not at all representative of their workflow or process.

- Don't keep banging your head against the wall.
I very rarely do batches. If you don't get the result you want within a few tries with prompt adjustments etc, then switch checkpoint model. Don't try to force it to do something it has not been trained to do.
The same goes for issues with the eyes etc. Either try a different ckpt or simply fix it after with inpainting etc.
You could of course use an extention like after detailer etc to give SD a helping hand.

Indeed, I might have gone a bit overboard with the 'shotgun method' of adding tons of tags and phrases to my prompts.The reason? I let GPT-4 create the prompts for me!

 

[Sepheyer]

Indeed, I might have gone a bit overboard with the 'shotgun method' of adding tons of tags and phrases to my prompts.The reason? I let GPT-4 create the prompts for me!

- ChatGPT bro, create a prompt for the Little Red Riding Hood porn film actress costume.
- It is important to respect the women's feelings and concerns, thus the director and the actress should work together to create outcome acceptable to all parties.
- ...

 

[Mr-Fox]

I'm moving a convo with Thalies to this thread since I think it could help others also.

" Hi Mr. Fox,
As someone currently exploring the capabilities of Stable Diffusion through Fooocus, I'm reaching a point where I'm considering whether to continue with this tool or change to A1111.
I’m also curious about any addons you’ve found particularly beneficial.
Thank you. "

I have no experience with Fooocus or other simplified Ui's so I can't help you with those.
I'll give you my opinion though. They are likely based on A1111 (my speculation) and only a "dumbed-down" version so you might as well use the real thing.
Even if this is not the case, A1111 is mainstream at this point and you can find much more information and guides about it, as well as extensions etc. It's really not that complicated.
Just start out simple with the basics and go on from there.
Install A1111 first and start familiarize yourself with it. Then the first 2 extensions you should get imo is controlnet and after detailer.
They might be included by default at this point though.
Sd Upscale Script for img2img (not ultimate upscale) if you wish to upscale rather than using hiresfix for some reason.
If you are using refiner, hiresfix doesn't work well so I upscale in img2img then instead.
If you want to swap faces for your own or celebrities etc, get reactor.

I recommend getting NMKD-Siax sampler and Ultrasharp also.

You must be registered to see the links

You must be registered to see the links

You can install A1111 two ways. Either get the install exe or if you are familiar with github do a git clone.

A1111 install guide Sebastian Kamph:

You must be registered to see the links

Using A1111 Sebastian Kamph:

You must be registered to see the links

 

[shakus-gravus]

Appreciate all the info shared in this thread. I am only just starting to mess with AI image generation. Is anyone aware of a good series of videos that unpack how to get a1111 all setup without focusing on every little nerd-knob and feature? Basically, I'm looking for an a1111 Deployment Guide along the lines of a Quick-Start guide that gets you up and running with everything you need as fast as possible so that you can spend more time creating than configuring the tool you're using to create ;-)