Giantess foot entrapment tests

In service of the bigger giantess feet LoRA I'm working on, I decided to separately train models for the different poses I want to capture, then eventually merge them into a larger general model with more control. I'm starting with sock/shoe/pantyhose entrapment. I might consider posting this one either as an independent lora or as a version of the original just because I think it works better than the general does.

It struggles a bit with foot anatomy without using the other foot LoRA as a sort of helper (which also means locking it into the style that the current version of that LoRA tends to push). However, it does do what it's supposed to do. Probably going to retrain the general LoRA with better style prompting so I can boost the success rate, touch up a bunch of sample images in a variety of styles, and give it another go.

These are all the raw output of that LoRA in its present form, so no hires/inpainting. Naturally foot anatomy is a challenge and it struggles to give the illusion of transparent fabric rendered in front of the subject (that will likely require a lot of manual editing to build a training base), but it is surprisingly effective at times.

(some of these prompts are a bit clunky as I procedurally generated them with wildcards and prompt reverse-lookup, and some of them may look familiar to other posts as I reuse some prompts as reference)