[Stable Diffusion] Prompt Sharing and Learning Thread

[Synalon]

forgive me I still don't understand what you mean by training... I only understood that Lora uses pre-existing models and XL uses Meta Human models

To make a lora you "train" it by giving it a bunch of pictures of what you want it to be about from every angle. It then spends hours repeated looking at them and trying to recreate them, (thats the training part). After that if it worked well enough for your tastes its done.

Its actually far more complicated but thats the easiest way I can think of to explain it.

 

[TitaniumDickDiamondBalls]

Does anyone here have defacounter_mix10 model ? It used to be on Civitai, but was taken down from the site for some reason.

 

[Rovente95]

how can I paint a image?

 

[Synalon]

how can I paint a image?

You mean inpaint?

Put the image in img to img, and theres a tab below called inpaint. tick that and draw over what you want to change, select resize and refill and generate I guess.

 

[Rovente95]

Thanks everyone, Now I need to study "training" to XL

 

[devilkkw]

A little simple test with artist i like: Gris Grimly
Prompt: woman by Gris Grimly

Result with no negative:

Spoiler: no negative

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

Result with

You must be registered to see the links

Spoiler: New-negative-v1.4

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

And one with

You must be registered to see the links

Spoiler: Easy photorealism

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

Image in png to check setting.

In general i don't like triggering artist, but for checking embedding or checkpoint i usually do simple test like it.
If i trigger an artist (with no negative) this do a 2d image, and getting something different like 3d or photo, i waste checkpoint.
This is a simple example, i'm currently testing checkpoint about all 2416 artist with this prompt.
Maybe i made a topic with all result if you interested.

You found all Artist on my

You must be registered to see the links

 

[me3]

For most people here SDXL most likely isn't gonna be "worth it", specially not if you're not already somewhat familiar with the basics.
Reasons being that very few things will work with it and the few things that does are still gonna be in a development and testing stage which creates a lot of oddities, bugs and other things that'll lead to results being different.
There won't be much "add ons" or embeddings of any kind compared to what you have with sd1.5 and as they show up they'll likely be less optimized, less accurate etc
A point that will apply more to newer users is that regardless of your system, you'll learn how prompts work, what doesn't etc much faster with sd1.5 than sdxl for the simple fact that XL requires a minimum size of 1024x1024, meaning you have to generate 4x the size each time you try to work out your prompt. Anything you learn about prompts will remain the same, basics are very unlikely to become that different and you'll have a much wider selection to play with, much more options in the results you can achieve and better fitted "tools".

 

[felldude]

I know I can't train on XLSD, maybe you could train with a 12GB card likely you need a 24GB.

And that is with optimizers, 40GB cards won't cut it even for LORA's if your trying to due full FP32 training.

Im guessing maybe 10% of the community will be releasing XLSD content.

I'm getting decent results using Realistic vision or another 1.5 checkpoint as a refiner. Average a minute for 1024x1024 keeping both checkpoints loaded in memory with Ergan4x+ as the upscaler.

 

[Mr-Fox]

Yes, for example their is a face set of data trained from Meta that is 100GB, (They tell everyone not to use it for facial recognition wink wink)
Most people can't do anything with it.

Training from scratch involves a 100k rig or more and usually takes days even with 80GB of video ram and 32 processors feeding that.

When us normal folk train a lora or checkpoint we are training with an existing set so in this case it would be SDXL, how it responds to training with 8bit optimizer or FP16, will determine the number of modified checkpoints or loras that support it

That sounds very interesting. Do you have any links to what you are talking about so I can do some research?

 

[felldude]

That sounds very interesting. Do you have any links to what you are talking about so I can do some research?

I think the trained set is behind the devolper wall but the

You must be registered to see the links

So it was Nvidia, Meta, and Amazon, looks like the Nvidia team was the main contributor for it.

I've never done the work to get dali, nvcc and all the other programs needed to use one of Nvidia's sets, they have some of the best in the world though...like the real time set that turns

You must be registered to see the links

The way their models work the CPU and GPU are used and coding is done on the fly, maybe some extremely talented person could convert their pickles to something usable in SD

I did find where they put out Stylegan2 in

You must be registered to see the links

What I find most interesting in that article is they claim Full FP32 training with only using 8.3GB of VRAM using 1024x1024 images....I dont have 48GB of System ram but still (Or a telsa V-100)

You must be registered to see the links

 

[namhoang909]

View attachment 2733142

Language Models

Want AI to write you porn stories? While being completely uncensored, self-guided, using desktop install, and running on a potatoo PC?

Here is the quickest no-frustration guide that will have you running right away:

1. You must be registered to see the links
   - go with "One-click installer" for your OS. Save the zip to the like of: "D:\LMs\oobabooga" and then run the install batch file. Once the install completes you will have an equivalent of A1111 webui installed in the folder from which the batch file was run.
2. You must be registered to see the links
   . Grab the "guanaco-7B.ggmlv3.q4_1.bin" file, that's the actual model, and save it inside "..\oobabooga\text-generation-webui\models".

That's all you need.

Housekeeping regarding settings:

1. Go to tab "Parameters", set "max_new_tokens" to the very max. As of today, the fresh install comes with max of 2000.
2. Go to tab "Chat settings", then subtab "Instruction Template", choose "guanaco non-chat".
3. Go to tab "Text generation", input field, and then: "Write a script for the red riding hood porn film." and watch the magic happen.

PS Also, a relevant video driving the point of why the tokens are important and the race for increasing the token counts:

You must be registered to see the links

PS2 Here is how the LM described a red riding hood outfit for a porn movie that resulted in the picture below:

Spoiler: LM Output

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

View attachment 2733228

Spoiler: More

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

So you use a text to text AI to create the prompt and use that to generate picture in text to image AI, is that right?

 

[Mr-Fox]

Floating Sorceress.. In colab with the eminent Synalon . Don't look too close she might bite..

 

[hanpengyi]

Can anyone share what is the latest cutting-edge NSFW model? I started playing around with stable diffusion today, but it takes years to generate an image (Apple M1 chip), and the online version is not free and doesn't allow NSFW content. I'm especially interested in models that are good at generating fake celebrity pictures, so preferably models with large tranining data on celebrities or have track record of producing good celeb fakes. Thanks.

 

[Mr-Fox]

Most SD1.5 models on civitai does NSFW content. I would recommend Clarity, because it is very consistent. There might be other checkpoint models that produce better images but very few are as consistent as Clarity. For the best result creating celeb fakes either use a Lora for this specific celeb or use the roop extension. With it you are using an image as reference and it is included in the last steps in the generative process. Thank you Synalon for telling me about this extension.

You must be registered to see the links

You must be registered to see the links

examples:

 

[felldude]

Well I did get a setup to train a lora on XLSD, loading latents from disk and using 640,640 instead of my usual 768,768....

10 hours for 1500 steps is about 100 times longer then I am used too though....Im guessing that would be 100 days to train a checkpoint on my machine

Even dropping to 256,256 will take 4 hours

 

[Sepheyer]

Trying to see what will give me Honey Select 2-like backgrounds, and stumbling upon some various pleasant looking pieces:

Spoiler: More

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

 

[Mr-Fox]

Well I did get a setup to train a lora on XLSD, loading latents from disk and using 640,640 instead of my usual 768,768....

10 hours for 1500 steps is about 100 times longer then I am used too though....Im guessing that would be 100 days to train a checkpoint on my machine

Even dropping to 256,256 will take 4 hours
View attachment 2816717

10 hours isn't as bad as you might think. My Kendra Lora took ap 16 hours with 3 epochs, 23 images, 768 res, 100 steps per image, 2300 steps in total per epoch, using slow learning rate and settings for "dampening".

You must be registered to see the links

Kendra Lora

You must be registered to see the links

 

[felldude]

10 hours isn't as bad as you might think. My Kendra Lora took ap 16 hours with 3 epochs, 23 images, 768 res, 100 steps per image, 2300 steps in total per epoch, using slow learning rate and settings for "dampening".

You must be registered to see the links

Kendra Lora

You must be registered to see the links

If my math is right you where training around 4.5 sec/it at 768,768 and I am at 30 sec/it at 640,640 and 7-10 seconds per it for 256,256 (It is stretch for me to be training antyhing on a 8GB 3050 at all though)

It is up and available for

You must be registered to see the links

 

[Mr-Fox]

If my math is right you where training around 4.5 sec/it at 768,768 and I am at 30 sec/it at 640,640 and 7-10 seconds per it for 256,256 (It is stretch for me to be training antyhing on a 8GB 3050 at all though)

It is up and available for

You must be registered to see the links

View attachment 2817177

I have only an old pascal card. GTX 1070 Gigabyte Overclocked 8GB Vram. So you should be able to at least match my numbers with a 3050. I'm just guessing, but it would be very strange if a card 2 generations younger isn't better..

 

[felldude]

I have only an old pascal card. GTX 1070 Gigabyte Overclocked 8GB Vram. So you should be able to at least match my numbers with a 3050. I'm just guessing, but it would be very strange if a card 2 generations younger isn't better..

Training on SD 1.5 at 768,768 I average 1 IT/Per second but XL is FP32 the math is exponential

Settings also play a role xformers, buckets etc