2023/6/17 update
Fix the problem of poor image quality generated by using WD VAE. The preview images are generated using 1.6 version, 2.0 version Yesmix and MeinaMix to verify the compatibility of this LoRA model on different base models.
2023/5/29 update
Update dataset. New version is trained using new craft.
Introduction
The original intention of creating the Fashion Girl model was to generate fashionable anime girl images (2.5d girl images). The idea of creating a new model emerged in order to generate more rebellious and cooler girl images. Therefore, this LoRA model can be used to generate cooler and naughtier girl images compared to the Fashion Girl model. The trigger tag is 'bad-girl'. The 784mb VAEs (NovelAI, Anythingv3, Orangemix, Counterfeit) are recommended. 0.6 ~ 0.8 weights are good. The preview images are generated using 1.6 version (original) YesMix.
This model will be continuously updated as the dataset is updated (although images that satisfy the conditions are not much).
Cautions (Must Read)
For anime model,
pls disable 'face restoration' during generating.
pls disable 'face restoration' during generating.
pls disable 'face restoration' during generating.
Some preview images are generated using 'Hires Fix' and upscaled using 'SD Upscale'. Thus, if you wanna generate the preview images, pls download the original preview images and analyze them using 'PNG INFO' in webui to get all generating information.
BTW, make sure set this option in 'Stable Diffusion' settings to 'CPU' to successfully regenerate the preview images with the same seed.
The training resolution of this model is 1024x1024, so generating resolution must not exceed this range. Try to use 'hires fix' to prevent this problem and help to generate higher-resolution images.
For 784mb VAEs (NovelAI, Anythingv3, Orangemix, Counterfeit) users, make sure add '--no-half-vae' in command line of webui to prevent generating black images. See voldy's log for the details of edit method. For high-performance GPU, make sure 'medvram' and 'lowvram' mode are disabled.
Description
Update dataset. New version is trained using new craft.