This is a Textual Inversion Embedding of Marin Kitagawa from My Dress-Up Darling (sono bisque doll wa koi wo suru).
I wasn't familiar with the character when she came in as a request, so I watched My Dress-Up Darling over the weekend, was immediately hooked, binged the whole thing, and now I'm pretty obsessed and can't wait for season 2! So I really wanted to get this right and tinkered with it way too much.
It's mostly trained on both her school uniform and iconic bikini, so you can add tags to steer output towards either, though if you don't do that you'll sometimes get a mix of the two. My initial attempts had the multiple ear piercings as a focus but this really upset SD generating her ear properly, so you'll have to add prompts for earrings/piercings to get these consistently now. Same goes for the fake/long nails if you want those.
Example images created using my 7th Heaven mix https://civarchive.com/models/4669/corneos-7th-heaven-mix
Check out my other civitai models for more characters and sex positions! https://civarchive.com/user/Master_Corneo
If you enjoy this embed, I genuinely love seeing your creations with it - you can attach them to a review if you're not aware (but not comments sadly)!
Description
FAQ
Comments (31)
good job, if you can please make textual character from genshin impact/ honkai impact
Anyone in particular? There's one on the linked poll already
Keep up the amazing work! Any chance you could do a hypno TI?
Probably, but can you give me a couple of examples of what you mean?
So this is probably far more examples than you wanted but hypnosis/mind control takes a lot of forms, and crosses over into a lot of other tags so I grabbed a bunch so you could get a wide scope and see the general trend/pattern. I hope this helps! Let me know if you need anymore info or have any questions.
https://pbs.twimg.com/media/FCEC-poXMAIC1MZ?format=jpg&name=small
https://hentaidevil.com/uploads/posts/cover/medium/46/468136-27925.jpg
https://static.hentai-img.com/upload/20140501/1/889/1.jpg
https://pbs.twimg.com/media/Ed247i3XgAAoSl8.jpg
https://freehentaipic.com/images/2022/10/19/hypno-tentacle-hentai-3.jpg
https://freehentaipic.com/images/2022/10/19/hypno-tentacle-hentai-2.jpg
https://i.kym-cdn.com/photos/images/facebook/002/077/763/82e.png
https://cdn.donmai.us/original/57/aa/57aaf1181db60fe51659e828290ab619.png
https://pbs.twimg.com/media/EfmJ0iOXYAY6f48.jpg
gif: https://i.imgur.com/O21L2t8.gif
https://i.imgur.com/F4Kqmol.jpeg
https://tse4.mm.bing.net/th?id=OIP.dyy6VUFz50xE9LiHaZvIxQAAAA&pid=15.1
http://pictures.hentai-foundry.com/r/RenaissanceOfChaos/296505.jpg
https://bowsettehentai.com/uploads/posts/cover/medium/192/1924174-9dc3a.webp
gif: https://sexdicted.com/wp-content/uploads/2019/07/hypnotized_mind-3206.gif
no futa please: https://chiks.org/uploads/posts/2022-09/1664027629_24-chiks-org-p-futa-hypnosis-erotika-32.jpg
https://e-hentai.org/s/4d814300c6/1050234-21
https://e-hentai.org/s/7d5ee43fcf/1050234-14
https://e-hentai.org/s/e5f929e544/1050234-1
https://danbooru.donmai.us/posts/3826784?q=mindflayer+
https://danbooru.donmai.us/posts/5958014?q=hypnosis
https://danbooru.donmai.us/posts/5884730?q=mind_control
https://danbooru.donmai.us/posts/5845918?q=mind_control
https://danbooru.donmai.us/posts/5770070?q=mind_control
https://danbooru.donmai.us/posts/5673293?q=mind_control
https://danbooru.donmai.us/posts/5640977?q=mind_control
https://danbooru.donmai.us/posts/5640992?q=mind_control
https://danbooru.donmai.us/posts/5627751?q=mind_control
https://danbooru.donmai.us/posts/5550515?q=mind_control
care to share your parameters ? (training steps/images, number of images, character pose, ratio face body, model used to train).
also thx for the release ! it looks stunning ! you got the hang of TI training and you became the standard to me.
As I mentioned I did tinker with this a lot, really appreciate the kind words! At the end of it all I still just follow this guide really https://rentry.org/simplified-embed-training
@Master_Corneo thank you!
Is it possible to train a character only with grayscale manga images and generate color images? I want to test it, but first I want to make sure it works.
I want to train Schierke from Berserk.
I'd say it's somewhat likely but I haven't tried it and if anything I go the opposite way, I always make sure not to use any greyscale images in my training data, but yeah if you're going to use stuff straight from the Berserk manga that might be your only option. As long as you leave stuff like black and white and greyscale in the tags alongside the training images, I think it'd work. Good luck! Let me know in fact!
@Master_Corneo I already tried and I think it didn't work well.
I did it with 24 images from the manga, most of them are of the face, some of the middle torso, all with their grayscale background. Would it be better to remove the backgrounds?
Learning rate:
0.05:10, 0.02:20, 0.01:60, 0.005:200, 0.002:500, 0.001:3000, 0.0005
@fatal3rror I follow this guide https://rentry.org/simplified-embed-training which suggests you need a lot more images
You are providing me with the TI I want, and I am so happy that I am making so much progress in creating images that I don't even have time to learn!
It's not even a LoRA :)
@Master_Corneo Sorry, I understand it is TI. I made a mistake. It is true that they are active in my environment. I am very grateful for your support.
@monototo Thanks for the kind words! I appreciate it
How to use It?
You put the .pt file in the embeddings folder, then use the filename without extension as part of your prompt (can see on the image metadata on the examples)
@Master_Corneo How do I see metadata?
@oiue There's an (i) icon on the images here, or you can drop them into the PNG info if you're using Automatic1111's web UI
@Master_Corneo what model should I use?
@sapple anything anime-based should work! I have two models here on civitai which both work well
@Master_Corneo one last thing, what prompt did you use?
@sapple it's in the metadata of the image, visible here as I explained above
Thanks
how come my outputs are only vaguely resembling marin and not as well generated as the examples?
Where can I download your NFSW models and text embbedings?
是一部很好的动漫，同样期待第二期
I really want this to work but its really weird. I sort of get Marin but its almost like a water painting when I render it. Im even copying the prompts and negative prompts from some of these images and dont get Marin half of the time.
hey ther mistr, could you pls share the link of video or site where you learned to train models this good?
Damn that Embedding works better than 90% of Lora's with her, I love it
Details
Files
corneo_marin_kitagawa.pt
Mirrors
corneo_marin_kitagawa.pt
5251_6093_corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
corneo_marin_kitagawa.pt
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.