- May 18, 2018
- 57
- 266
- 238
Alright, download link is up. Sorry again about the delay. Let me know if anything seems broken.
Stable Diffusion on ForgeUI. The model will depend on the picture. Version 0.1 and most of 0.15 were made on a ten year old (checks calendar year, jfc) 13 year old computer, so due to graphics card constraints most images of people were made with an SD1.5 model (usually Mistoon Saphire v20, made by Inzaniak on Civitai) then touched up by hitting it with ModelEX_v60 (I think the author, Nico53425342 on Civitai, replaced it with v70 on civitai). Backgrounds and objects were usually the SD1.5 Arthemy Comics, now ArthemyComicsXL after the computer upgrade. Then a lot of editing with gimp, then throwing it back into the img2img, things like that. I'm still using SD1.5 Mistoon Saphire even after the computer upgrade though, its where the unique (for AI images, at least) faces come from.What AI generator do you use
I'd suggest pixai.art if you're looking to improve on models- not only does it have all the standard models but it has user submitted ones as well- and some do things better than others- Haruka v2 as a model is very good for r18 and stays on model if you use a base.. You can also train the model with free points- with a LoRA- this lets you set it up for specific styles and specific actions- for instance there are some usermade ones that would make Haruka v2 excel at open-door NTR scenes- or a specific artist style(I know it's in bad taste but many are simply trained by uploading dozens of the same artist-but the outcome is actually impressive).. Main thing is that it's free- you can create as many accounts as you need- no email verification is required but you get a discount on generations if you do link it- you can also activate some bonus claims(like watching a video- following on discord etc- though you dont actually have to.. just have to click the link) which gives enough to create the LoRA when normally you'd have to wait 2 days of daily points to get enough. LoRA aside though- I think you shouldn't need it and just be fine with a base image on most of the models provided. It does really well and i'm just surprised none of these AI games ever bother to use it(I can tell from the style they're not made with those models)Stable Diffusion on ForgeUI. The model will depend on the picture. Version 0.1 and most of 0.15 were made on a ten year old (checks calendar year, jfc) 13 year old computer, so due to graphics card constraints most images of people were made with an SD1.5 model (usually Mistoon Saphire v20, made by Inzaniak on Civitai) then touched up by hitting it with ModelEX_v60 (I think the author, Nico53425342 on Civitai, replaced it with v70 on civitai). Backgrounds and objects were usually the SD1.5 Arthemy Comics, now ArthemyComicsXL after the computer upgrade. Then a lot of editing with gimp, then throwing it back into the img2img, things like that. I'm still using SD1.5 Mistoon Saphire even after the computer upgrade though, its where the unique (for AI images, at least) faces come from.
I've downloaded both and both are the same size when downloaded must just be a visual thing for mega.The file sizes are different for Gofile vs Mega?
*edit* - to clarify, I'm not suspecting foul play, just curious if a different version got accidentally uploaded or something or if it's an idiosyncrasy with the sites.
Double checked this, since I only put the mega link up. The gofile size ends up being the same size as my original copy of v0.15 once everything is extracted.The file sizes are different for Gofile vs Mega?
*edit* - to clarify, I'm not suspecting foul play, just curious if a different version got accidentally uploaded or something or if it's an idiosyncrasy with the sites.
Thanks, just thought it was odd.Double checked this, since I only put the mega link up. The gofile size ends up being the same size as my original copy of v0.15 once everything is extracted.
They're both in the game.Are the scenes in the last two images in the game yet? Also i tried few times but can't find the ghost in the 2° hunt, is there just a bad ending for now?
I just got the second to last image. I think you have to have a bit of a relationship with Noms. The last pic is probably at the same time but when you build your relationship to queeny instead. I think. Based on what she wears and the way she puts her butt apparently right in front of your face.Are the scenes in the last two images in the game yet? Also i tried few times but can't find the ghost in the 2° hunt, is there just a bad ending for now?
Yep! And my hope is to balance things so it won't require 'pushing' the girls towards something, it'll just happen naturally if the player isn't carefully balancing all of the other ghost hunting mechanics properly.So if you fail the hunt, it's NTR? And if you succeed, it's not? Unless you kind of push the girls to doing so in the haunted houses?
I like games where NTR is a threat but you can avoid it by playing your cards right, so the game sounds up my alley.
Tried it game a bit, was kinda fun but I think I need to go over the tutorial again lol, guessing the right ghost is hard!
No definite plans. I'll play around with AI animation eventually, since I think the camera scenes would be better animated. I don't think I'll try my hand at any more amateur semi-animation though lolAlso mr Dev are you planning on adding animations? I know some scenes are "semi-animated".