Search results

I would also be interested in tour workflow. Which base model and, if any, which lora do you use for these? ___ Also, I will be posting some more 'wet dreams' soon :D

A few more edibles:

Some doodles:

A few of my doodles.

Woah! God tier rendering, mate.

More tweaking of my stand-alone Checkpoint. I've completely given up on attempting 'photorealism' at this point. The 'semi-realistic' style has always appealed to me more, so I'm embracing it fully from this point onward.

More of that inner freak :P

Sometime you jus have to throw realism out of the window and just embrace your inner freak:

Spent the day tinkering with fabrics again. There is naturally enough, one particular type of fabric that appeals to me, can you guess which? :D

Ah, the holy grail of NSFW AI. NSFW AI Image to Video, (paid or free to use) is something I've also been looking into for a while now. While there are definite signs that it is coming, I don't think we'll see them just yet. (I would however be very happy to be proven wrong). Putting together...

The original image was rendered and then animated on a site. I'm still looking for a method of doing it locally.

While I am first and foremost a fullback panties guy there is something to be said when it comes to a pretty lass in a thong: (All created with the latest version of my Checkpoint)

New lora (above) merged into my base\checkpoint model.

Testing a new lora. This was trained with 10 passes on 220 images. The base model is, once again, FantasmPonyXL.

This is checkpoint, a StableDiffusion model (Fantasm) with several of my lora merged into it. The lora were each trained on around 100 images each. There 7 or 8 lora in this particular model.

When your PTSD is kicking your ass, make something pretty to cheer yourself up:

Aya <3 - Thank you for the re-up, my collection is almost complete thanks to this.

Tweaking my Checkpoint toward a more realistic look:

@Casshern2 - Had a play with your prompts :) My 'PLP' model was trained on 20-30 somethings so it has a bit of trouble with the age, but the prompts seem to do a lot of heavy lifting. (Still undecided as to whether or not I need more realism in this model). Welcome to the homebrew-rendering...

Not exactly 'online' but I use kohya_ss for creating the initial prompts for the models I've been training. It's a relatively quick and easy install, and can generate prompts from literally thousands of images in a couple of minutes.

More improvements to my standalone model. Still has the 'A.I generated' look and feel but a vast improvement on things like hands and body proportions, and far better adherence to prompts.

;)

So, who's been naughty?

My current obsession, Tsubomi Mochizuki:

New year, new version, new model, new girl: _

Torn between 'just a liitle bit late, 2024' and 'too damn soon, 2025', but Project Lovepop just hit a new landmark with version 05.1 so have chistmassy pic:

There's been an accident in the kitchen! The accident:

@Casshern2 Something like this? The prompts are very simple, (auto generated by the site I used', but can be further improved, if needed, to better represent the original, i.e. 'large breasts' to improve bust size, 'thick thighs' for bigger thighs, etc. (I tried to render a copy with...

Latest iteration of my 'Project Lovepop' model. Slowly but surely, it's getting there. This is a composite of around six to eight-ish Lovepop girls. I've been a kind of Baron V Frankenstein for the last few days, building my ideal woman from bits of other lasses :D __ Hands are...

By the looks of these I'm assuming you'd like to take a sketch, or similar, and make it more 'real'? I'm no expert, but I may be able to find a method for you. At work right now so I'll take a look in a few hours and get back to you.

Your Christmas present is here, you just need to unwrap her :) (I'm sold on Forge, what a great toy!)

Ready to be unwrapped and eaten:

Also: Merry Christmas

Oh my! Where do I find the photoset that the Kan Yura picture comes from?

I've just made the switch from EasyDiffusion to Forge. The difference in quality is very, nice. A lot more to learn, but Forge seems to have it all.

A quick lunchtime doodle. Still not as 'real' as I'd like, but another step in the right direction.

Further progress with Project Lovepop: