Another new day a new resumption

[JSample]

Hey, cerberus, you've probably already answered this question elsewhere for someone, but which AI image rendering software do you use? I've been cutting my teeth on Microsoft Designer, and while I'm generally pleased with its results, I am running into issues and inconsistencies with what I call its "prurience monitor." For example, from experimentation I've learned that Designer will readily generate images of women sunk in quicksand so very deeply that only their eyes and forehead are visible above the surface (and sometimes as deeply as only leaving the tops of their heads and their hair visible), but for some reason it balks at showing women "standing" knee-deep or waist-deep in a mud pit or in quicksand ("Your prompt may contain wording that violates Designer's code of conduct.") I just thought I'd ask to see what other options I might have. Thanks for all your inspiring work here.

[cerberus]

Hey, cerberus, you've probably already answered this question elsewhere for someone, but which AI image rendering software do you use? I've been cutting my teeth on Microsoft Designer, and while I'm generally pleased with its results, I am running into issues and inconsistencies with what I call its "prurience monitor." For example, from experimentation I've learned that Designer will readily generate images of women sunk in quicksand so very deeply that only their eyes and forehead are visible above the surface (and sometimes as deeply as only leaving the tops of their heads and their hair visible), but for some reason it balks at showing women "standing" knee-deep or waist-deep in a mud pit or in quicksand ("Your prompt may contain wording that violates Designer's code of conduct.") I just thought I'd ask to see what other options I might have. Thanks for all your inspiring work here.

I use a locally installed version of Stable Diffusion(SD) front ended by Automatic1111(A1111), by this I mean I have the software installed on my home PC and am not using an online service. Stable Diffusion is the AI generator behind a number of online services (e.g. Perchance, MageSpace etc), it is Opensource and is free to use. Automatic1111 is a local browser app (again not an online page) that provides facilities to allow the user to select and set all the features and options required by SD.

Using this has a number of pros and cons. I'll deal with the cons first:

You need a good spec machine with an up to date Nvidia or AMD graphics card (Nvidia preferably) and 8GB graphics memory absolute minimum.
You need a lot of spare disk space, first day of downloading I'd filled 30GB, lots more was to follow.
It's not trivial to set up.
It is a steep and extensive learning curve.
There's help online but it's piecemeal and specific to particular topics.
The range and scope of additional options and files available is bewildering.
It's not perfect, you'll get excess fingers, toes, hands, arms, legs, backwards feet and hands, etc, so you need to be good with Photoshop etc. Some will be OK though.
People, who should know better, will tell you AI is not art, it's so easy to put in a prompt and spam as many pictures as you want.

The pros:

It's free.
It's pretty good.
There are loads of checkpoints, lora, etc available for free.
It is largely uncensored, it does seem reluctant to do any violence.
It appears to only be limited by your imagination.
It has a sense of humour.

[cerberus]

Saturday 1.

00002-2327840427.jpg

00031-2892533521.jpg

00066-857709103.jpg

00074-2059674301.jpg

00076-1236738459.jpg

[cerberus]

Saturday 2.

00082-4059312806.jpg

00111-1519117273.jpg

00140-313212861.jpg

00152-3053453990.jpg

00235-3710259958.jpg

[JSample]

I use a locally installed version of Stable Diffusion(SD) front ended by Automatic1111(A1111), by this I mean I have the software installed on my home PC and am not using an online service...

Thanks for the info, cerberus. I've got enough free space (2.7 TB out of 8.2 TB total), but my 4-year-old Nvidia card has only 4GB RAM, so until I upgrade that I guess I'll be sticking with Designer.

[MadMax359]

excellent ropework on the first damsel!!

[cerberus]

I use a locally installed version of Stable Diffusion(SD) front ended by Automatic1111(A1111), by this I mean I have the software installed on my home PC and am not using an online service...

Thanks for the info, cerberus. I've got enough free space (2.7 TB out of 8.2 TB total), but my 4-year-old Nvidia card has only 4GB RAM, so until I upgrade that I guess I'll be sticking with Designer.

If you ever decide to give it a go, I'm more than willing to answer questions and give you a helping hand where I can, so feel free to ask. 2.7TB may be enough, it depends how addicted you get! my folder is now at about 400GB most of which is models (checkpoint, LORAs etc), each SDLX checkpoint file is over 6GB (I've got about 50 of them), I'm only keeping about a months images in this folder. I'm not seeing too much development of SD at the moment, so it is possible it may be superseded or fall into disuse, the company that produced seems to have been in a bit of turmoil recently.

[cerberus]

excellent ropework on the first damsel!!

It's a bit difficult to tell under all that mud. Still, looks like a fun day was had by all!

[cerberus]

Sunday 1.

00020-3306090357.jpg

00035-592862533.jpg

00085-1511309830.jpg

00093-2945779698.jpg

00138-1615374350.jpg

[cerberus]

Sunday 2.

00186-2142175987.jpg

00260-546357527.jpg

00282-2342172681.jpg

00310-3156522582.jpg

00311-1411522093.jpg