Lily from Spy Kyoushitsu (Spy Classroom)
Prompts:-
Trigger Word:
lily_olisBlack Outfit:
black shirt, neck ribbon, suspenders, high-waist skirt, black skirt, black pantyhoseExtra:
1girl, solo
black footwear (may not resemble her boots much)
Weight: 1 (I haven't tried other values)
My first model.
I've been experimenting with a lot of stuff over the past month, but I haven't been able to make any further improvements at the moment. As such, I've decided to share the model here.
Hopefully some of you will like it.
Description
FAQ
Comments (5)
Do you have a discord by any chance? I'd like to ask you a few things about how you went about making this multiple-outfitted LoRA, as I'm trying to go about making my own. Any help would be appreciated!
Hello. Thanks for your question.
I have a discord but I rarely use it.
To answer your question, I am afraid you have misunderstood something. In fact, only one outfit is included in this LoRA. The other outfits come from the base model, e.g. AbyssOrangeMix2. You can find the base model used on the right side next to "Model" after clicking on the images above.
I tried to include other outfits in this LoRA at first but the results were not satisfactory. You can ask those who have done multiple-outfit LoRA for help.
i'm no professional, but im pretty sure all you need to do is tag each outfit while you're setting up your data, and then it'll just naturally learn the girl + her outfits
for example.... idk madoka
tag her magical girl outfit with 'Magical girl' and her regular outfit with like 'regular outfit', and by the time its done training, you should be able to use Madoka herself for any outfit, 'magical girl' for her magical girl outfit, and 'regular outfit' for her regular outfit.
idk though i dont train loras i could be spewing nonesense. IF I AM RIGHT, this is also how people train multi-girl LORAs. except replace outfits with names
idk what im doing wrong, but her skin is always grey. Ive added like (((grey skin))) to the negatives but to no avail. also tried (((white skin))) to the positive but idk.
@NotTheNax
Hello,
Did you forget to specify a VAE?
You can add the VAE menu to the top of your webui by following this:
https://www.reddit.com/r/StableDiffusion/comments/11wjow6/quicksettings_toolbar_for_auto1111_model_vae_lora/
For VAE, you can try these:
https://huggingface.co/stabilityai/sd-vae-ft-mse-original/blob/main/vae-ft-mse-840000-ema-pruned.ckpt
https://civitai.com/models/22354/clearvae
https://huggingface.co/NoCrypt/blessed_vae
https://huggingface.co/hakurei/waifu-diffusion-v1-4/tree/main/vae
Put the downloaded VAE into the "VAE" folder under "stable-diffusion-webui\models\"
More info: https://rentry.org/sdvae
Also, try putting (greyscale, monochrome:1.1) to your negatives.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.