~ This is AI generated (with love) ~

--------------------------------------
A mage wanders off into the desert in search of her lover, when out of desperation she is forced to cast a survival spell she isn't quite ready for...
--------------------------------------

I put in some more effort into a single gallery this time, with a story. It's my first story, so please let me know what you think!

Also please let me know if I should continue to make stories. It took me about twice as long to do captioned, but IMO it was more fun.

Last edited on 03 June 2023, 05:06.

Base +6, covs -6, Qwertius +19, Xurantar +6, Gamefriend +10, wocoporn +6, DarthBogard -6, Spoot321 +6, randomguy5682 +6, Beowulfal +7, Tsilkani -7, loopyb +6, albertzeb -6, and 11 more...

I liked the concept! Please, do more like this story! You have my support.

Base +6, Xurantar +6, Gamefriend +10, Spoot321 +6, Adorable_War_7087 +2, Qwertius +19, loopyb +6, albertzeb -6, magicalstring -6, Shimaso1208 +3, fat_boy_tim5 +7, Flickr +6, kaguyasama00 +6, and 2 more...

OK, I have to admit AI is good at this mind-blowing and imaginative content.

Base +6, Xurantar +6, Saberling -6, DarthBogard -6, Supyeet -6, linksbro1 +12, Tsilkani -7, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6, Hardlizard +6, Laser-Brain +6, and 1 more...

uploader, can you explain a bit about how you make the images morph while remaining consistent please? Do you start initially with txt2img, then img2img for the rest of the sequence? How do you modify the lora amounts as you do so?

Base +6, DarthBogard +6, albertzeb -6, angrytaco -6, WillowRaison +5, Sherlockian +7, Laser-Brain +6

There is a spellcrafter making bank off of this booby-trapped spell. Sell the secret of how to cast the spell, then collect the resulting ready-made non-human slaves.

Base +21, albertzeb -6, angrytaco -6, Laser-Brain +6

it's amazing <3
I also would like a "how to", more to know what prompt and models you used :3

Base +6, DarthBogard +6, Higers -7, loopyb +6, albertzeb -6, angrytaco -6, magicalstring -6

maybe i should do this too

Base +6, DarthBogard +6, loopyb +6, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6, Laser-Brain +6

@sallazaar I start with txt2img and then move to img2img. To keep things consistent, I use controlnet, usually reference only, balanced for both txt2img and img2img.

Generally the flow is:
* Generate a bunch of test images at 512x768, img2img the results at 1024x1536, pick out one I like
* Use the same prompt with txt2img and turn on controlnet reference only mode, tuning the strength until I get a consistent character (usually 1.0 txt2img and 0.7 img2img, could be higher if clothing is more intricate)
* Run through a change to the prompt, adding in a little bit of loras, different directions each time (ex: kneeling down, hands on head, (lizard girl, scales: 0.4), etc)
* If the prompt is getting too fussy to accept a change, turn down the controlnet strength or use a more recent image as the controlnet base (I like to keep the original as long as possible to prevent small changes from occurring over time)
* As I get closer to the end, controlnet is usually around .5 strength

-----

@Morpheistos one of the later prompts is:

(ultra realistic, highly detailed:1.4), best quality, masterpiece,

girl at the desert, long blue hair, (blue skin)

((anthro):0.2)
((lizard girl, reptile, scales, scaled skin):1.1)
<lora:race-reptile:0.8>

((huge breasts))
((laying on back, cum, orgasm))

<lora:egg-laying:1>
Negative prompt: easynegative, bad-hands-5, (low quality, worst quality:1.3), child, boring_e621, naked

Steps: 25, Sampler: DPM++ 2M SDE Karras, CFG scale: 7, Size: 512x768, Model hash: f5ca7435bb, Model: transformix-v1, Clip skip: 2, ControlNet: "preprocessor: reference_adain+attn, model: None, weight: 0.8, starting/ending: (0, 1), resize mode: Crop and Resize, pixel perfect: False, control mode: Balanced, preprocessor params: (512, 0.5, 64)", Lora hashes: "race-reptile: 4db970e14abb, egg-laying: 2bfb26ef26c2", Version: v1.3.0

Model is https://civitai.com/models/83138/transformix
Egg Laying LoRA: https://civitai.com/models/44548/egg-laying
Reptile LoRA is custom. I'll try to get it on civitai soon.

----

@delta470 please do! It's a lot of fun and it would be awesome to see what everyone comes up with

Base +6, Qwertius +19, SxSov +6, Tsilkani -7, Nuuum99 +7, repeatedmeme +10, loopyb +6, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6, WillowRaison +5, Sherlockian +7, and 4 more...

yeah i already did that with some feminization story, i am just hating the waiting time for each generated batch. For now i run it on laptop with nvidia1650 which need more vram. If i run the webui then i can only wait and cant do anything else since it will crash if a little bit vram used on others program

Base +6, albertzeb -6, angrytaco -6, Laser-Brain +6

it was great till pg15, then the face started getting messed up, particularly with the horns and mouth. still would love to see more like this. thanks for adding the dialog as well, AI art is much better with a story.

Base +7, Gamefriend +10, Nuuum99 +7, Ulfhedinn +6, loopyb +6, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6, Laser-Brain +6, Regavachi +6

@uploader thanks so much!! Really appreciate you going into detail with the explanation, I think the community here can come up with some really intriguing tf ideas in the near future. Also, great work on your part with the investigation and creativity, this is an awesome imageset and story!

Base +6, loopyb +6, albertzeb -6, Polymorphed +6, angrytaco -6, Laser-Brain +6, Regavachi +6

missed opportunity for a coiling long tongue :O

Base +6, SxSov +6, loopyb +6, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6

and this is what AI art should be. Not 5k randomly generated images getting dumped in an untagged gallery.

Base +6, loopyb +6, keal6895 -6, fierylion +6, albertzeb -6, Polymorphed +6, angrytaco -6, magicalstring -6, WillowRaison +5, bobbycracky2 +7, fat_boy_tim5 +7, Laser-Brain +6, Regavachi +6

what AI Generated] you use o.o

Base +6

This is incredible. I usually roll my eyes when I see the AI tag, but I like the application of transformation to it.

Please do more. Captioned or not, it's great stuff. Would love to see like a doll or joker transformation, but all your stuff has been great so far.

I also see you disowned your other gallery, but you should be proud!

Base +6, loopyb +6, Laser-Brain +6, fat_boy_tim5 +7, Regavachi +6