I absolutely love Discipline Clinic. I'm a woman, and I saw you wanted feedback at some point from a female perspective! I just want to also say thank you for being inclusive of genders <3

Hmmm, looks like I have a call somewhere to "setImageOtkScold" that should be "set ImageOtk". I'll hunt it down. If you happen to see exactly where in the script this is happening, please let me know.

Hi.

Yes, things like not promising to not repeat an infraction can increase your time in some part of the script and some things will cause the mistresses to ignore your max corner time but usually they will tell you this when they send you to the corner.

Regarding coin training:

Thank you for your kind words! I'm doing my best by including all genders but it is difficult. I'll definitely keep doing it for DC, but some of my other scripts would nearly double the time it takes me to produce. It requires special code each and every time I use a gender specific phrase or body part or clothing. If there is anyone out there willing to help translate BDSM scenes from a male sub to a female sub, let me know so I can work that into "after dark"

I'm pretty sure this is what LoRAs are for. You basically fine-tune an image generation model on a set of 10-20 photos of a person. You can do the same thing with the outfit.

The output is then a hyper-specific image-generation model that should only output one model. With a corresponding outfit LoRA you can also fix an outfit. This should make generating the same person in the same outfit in different poses more reliable. Not saying each image will be good, but the hit rate should get high enough with a bit of tweaking to get usable results in an hour or so of fiddling around (after the intial cost of setup). Just a gut feeling tho tbh.

I'll take a looksie at some point over the next couple weeks, see if I can generate a moderately sized set of consistent images (~10 images) of a lady holding different spanking implements + scolding.

E.g: https://old.reddit.com/r/StableDiffusio ... onsistent/ - should be the same process for realistic characters. One probably doesn't even need to train LoRAs locally, you can just use stuff people have already trained.

I think there is a bug with the masturbation demerits and multi-sessions.

If you report masturbating without permission, and then split up your session over multiple subsessions, then you get a masturbation demerit for each subsession. I doubt this is intentional, since opting for a single session then likely only gives one demerit.

I'm guessing what happens is that some point loadOffenseData gets called in a subsession, then extraPunishmentMasturbation gets set in the case starting at 2232, and it triggers the demerit stacking in checkExtraPunishment.