https://huggingface.co/x1101/models/resolve/main/anymix_v2.ckpt (private)
Recommend Settings :
Clip Skip : 2
Sampler : DPM++ 2M Karras or DPM++ SDE Karras
Steps : 28
CFG Scale : 7
VAE: nai vae / blessed vae 2
This model is anything5 merged with mixed other nsfw model, so it can produce nsfw while retaining the clear flat anime-style,
Highly compatible with LoRA especially if you training it with anime-screencap dataset and use NAI model as base training.
until now i use it to generate nsfw stuff on my pixiv page. https://pixiv.me/x1101_id to see more example. all example image using this model
Description
different method and better result
FAQ
Comments (14)
Mind sharing the LoRAs you’re using on pixiv? xD
They all look so good ^^
sorry its personal , but alternatively you can use this guy lora https://civitai.com/models/4982/anime-screencap-style-lora , its similar with my lora, the different is just trained with ecchi and hentai image.
@x1101 I understand, thanks :)
Will the author still update this Stable Diffusion?
if there is an update on the base model, i will update this surely
@x1101 Can this Stable Diffusion make male characters?
@black_butterfly yes
@x1101 May I ask why many pictures are nsfw when making pictures?
@black_butterfly thats the purpose i make this things, model itself for nsfw purpose, you can put "nsfw" on the ngative tho
Is there any difference between AnyMix2_half and AnyMix2?
@black_butterfly more strong nsfw representation, the first one was a bit weaker