A mix of some of my favorite anime models so far. Geared toward character portraits for the most part, but is still quite capable of things like cars, landscapes and what have you. And, of course, porn, as far as more simple positions and concepts are concerned (for more complex stuff, you gotta use LoRAs). Rule 34 is as unbreakable as the law of conservation of energy.
Has issues with horizontally positioned characters' anatomy in landscape aspect ratios, but so did the original models, so yeah. Suggest models I could blend in with this to fix the issue if you feel like it.
Consists of the following models:
Eimis Anime Diffusion: https://huggingface.co/eimiss/EimisAnimeDiffusion_1.0v
YesMix: https://civarchive.com/models/9139/yesmix
Night Sky Yozora Style model: https://civarchive.com/models/12262/night-sky-yozora-style-model
The hypervenus bodies of the girls in the example pictures come not from the model itself, but from the LoRA I used: https://civarchive.com/models/16928/hyperfusion-100k-images
No VAE is baked in, the standard SD 1.5 one used for example pictures.
Description
Mixed in the Night Sky Yozora Style model: https://civitai.com/models/12262/night-sky-yozora-style-model
FAQ
Comments (4)
Looks very promising! but unfortunately I got runtime error when i load the M2 version. self.size(-1) must be divisible by 4 to view Byte as Float (different element sizes), but got 2367511
What UI are you trying to use this on? It works just fine on my setup of Automatic1111's web UI. Some UIs might be incompatible with the safetensors format, maybe that's your issue.
Automatic1111, but might be out of date. I'll try to update and try again, wish me luck
Working properly now, thanks
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.