Hugging Face's logo

hollowstrawberry
/
holotard

Text-to-Image
English
stable-diffusion
vtuber
hololive
stable diffusion 1.5
textual-inversion
lora
character
Model card Files Community
4
holotard / README.md
hollowstrawberry's picture
hollowstrawberry
Update README.md
e02ae40
preview code
|
raw
history blame
12.5 kB
metadata
license: creativeml-openrail-m
language:
  - en
pretty_name: Hololive vtuber LoRAs and TIs
task_categories:
  - question-answering
tags:
  - stable-diffusion
  - vtuber
  - hololive
  - stable diffusion 1.5
  - textual-inversion
  - lora
  - character
  - text-to-image
pipeline_tag: text-to-image
metrics:
  - character

Preamble

These resources are intended to be used with stable-diffusion-webui. If you don't know what it is or how to use it effectively, here's an extensive guide about it.

Recommended additional resources:

  • orangemix.vae.pt - without a VAE selected in the settings your colors will look faded.
  • TagComplete to view all available anime tags as you write your prompt.
  • Image Browser (updated fork) to view your past images and their prompt metadata.
  • Lycoris extension to be able to use locons and lohas, which are new types of lora.
  • EasyNegative - put it in your stable-diffusion-webui/embeddings folder and write EasyNegative in your negative prompt to drastically improve your images.

 

Models

These were mixed with hll, a model trained on tens of thousands of vtuber images. The 3.1 version knows more non-hololive vtubers, but has a different style. Many more mixes can be found here.

Click here for a comparison

A comparison of different vtuber checkpoints

 

Loras

All known holo Loras will be linked to and backed up here. Put them in your stable-diffusion-webui/models/Lora folder. Use them in your prompt like this: <lora:filename:0.7>
Most Loras work best with a weight of 0.7, but some recent ones work well at the default of 1.

Locons and Lohas are new types of Lora, they work the exact same but you need this extension to use them.

Learn to make your own Loras with my guide.

 

All Vtubers

These can do a wide variety of vtubers with only 1 Lora, but may have lower quality than a single Lora for each character. For a compromise between size and quality, try the Small Loras below.

 

Small Loras

These were scaled down from 144 MB to 18 MB or 36 MB and perform almost the same. They need a higher weight than the original to preserve detail. If one performs poorly for you, try the original instead.

This works because dim 128 used to be the default setting for making Loras but that was completely overkill and dim 8/16/32 work just as well.

This was done by me. Find them here.

Click here for a comparison

A comparison between a vtuber made with the original Lora and the compressed Lora A comparison between a vtuber made with the original Lora and the compressed Lora A comparison between a vtuber made with the original Lora and the compressed Lora A comparison between a vtuber made with the original Lora and the compressed Lora A comparison between a vtuber made with the original Lora and the compressed Lora A comparison between a vtuber made with the original Lora and the compressed Lora

 

Outfits

Main purpose of this repo. Format: (Source 1) (Source 2)... where each source can have multiple outfits.

The default outfit is the first/main one for that talent. Generic means it's not trained on a particular outfit and can do various clothing. In any case, most Loras can change clothes.

Example usage of each outfit can be found in the metadata of the image previews or in the text files next to each Lora.

HoloEN

HoloID

HoloJP

Staff

 

(Legacy) Embeddings

TIs or textual inversion embeddings are an older but lighter alternative. They go into your stable-diffusion-webui/embeddings folder and are used like a normal word in your prompt.

Nobody makes these anymore because Loras are easier to make and work better for most applications.