You don't have permission to view the spoiler content.
Log in or register now.
I use two generators:
1) available on civitai.com, link to model:
You must be registered to see the links
2) on home PC with automatic1111. Guide how to install:
You must be registered to see the links
- Jun 19, 2018
- 17
- 26
Feel free to give suggestions on characters, positions etc etc, I want to test the limits
- Sep 1, 2016
- 49
- 128
Damn some of these look a lot like the game. Devs could just train an AI with game content and release more updates.. Artist could then just edit the faces. I mean, Curse of Pleasure did it pretty well.
Can someone capture Ylva from game?
I don't have save with her appearance.
Also, looks like second model is able to generate Kira, but it hard (needs a lot of tries). Better than nothing)
And here some examples of groping and paizuri.
You don't have permission to view the spoiler content.
Log in or register now.
Also, looks like second model is able to generate Kira, but it hard (needs a lot of tries). Better than nothing)
You don't have permission to view the spoiler content.
Log in or register now.
And here some examples of groping and paizuri.
You don't have permission to view the spoiler content.
Log in or register now.
- Jun 19, 2018
- 17
- 26
I've been trying to get a proper Ylva, but its pretty hard, especially when it comes to her teeth. Although I made a few pics so maybe it helps somehow. Enjoy!
It is better not to use AI generated art, to train AI model, it will cause degradation.
You must be registered to see the links
Vehlis x Sabia lesbian 69, scissoring etc. Maybe Neve and Kira/Vehlis spit roasting or dual fucking Sabia with strapons?
I can work fine, but only if you edit it to fix any AI issues. Feeding raw AI art into a model will always cause issues though.
I mean no offense when I write this, but I do wonder where you guys are trying to go with the AI art based on this game. I mean I would love some patch, or reworked art for the game, because I do feel like this game has lost some quality over the past few updates.
- Jun 19, 2018
- 17
- 26
Personally I'm not going to make any reworks of the game or anything, just doing it for fun and if people like it, that's a nice bonus
I just trying to make usefull LoRA model, for some characters from this game.) Now i'm working on next version, and planning not to rush this time.
- Jun 9, 2019
- 242
- 927
Thanks for the LoRa my man!
Damn, that's some incredible work! I dunno if you're taking requests or not, but some of those look like you've done variants of real scenes. Would you be willing to try a Sabia version of the blowbang, or a horse version of the hellhound?
Some scenes easy to reproduce, because the have unique descripnion, others not. I try to fix problem in next version.
Both scenes included in training data, last one not with hellhounds, but with orcs.
First one is easy to invoke, at least with Elmy.
You don't have permission to view the spoiler content.
Log in or register now.
And second one not)
- Jun 9, 2019
- 242
- 927
that looks pretty good , can you direct me to some good tutorials i would like to try some as well
- Jun 9, 2019
- 242
- 927
I pretty much learned "on the go" and by all mean not a "expert"
Here you can learn the basic :
You must be registered to see the links
I also recommend creating an account on civitai.com once you have everything set up.
There, you'll likely find all the models (checkpoints), LoRa, etc., that you need. A great feature of this site is that clicking on an image provides you with all the necessary information: the model used, the prompt, the seed, the LoRa, and more. You can even use the "buzz," the site's currency, to remix and experiment with images using your own prompts, models, LoRa, and so on.
Also, despite its reputation, reddit is your friend if you have errors or some difficulties. Just tip on google whatever problem you have + reddit.
And last thing, I hope you have a good GPU
In addition to what was written by Mamaragan.
I have RTX 3070 8Gb VRAM. 1 image generation tooks 2-5 minutes, sometimes more. On site generator (civitai.com) is faster, generates 4 image per task, sometimes fails, but you get autorefund in this case.
Also, if copy all setting for image from civitai.com, it doesn't mean you get same result, it depends of GPU, noise generation and e.t.c... But you still get +- good result.
And with automatic1111 you can do more than on civitai.com (only txt2img).