|
--- |
|
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 |
|
--- |
|
|
|
<img src="https://huggingface.co/hollowstrawberry/holotard/resolve/main/out.jpg" width="800"/> |
|
|
|
# Preamble |
|
|
|
These resources are intended to be used with [stable-diffusion-webui](https://github.com/AUTOMATIC1111/stable-diffusion-webui/). If you don't know what it is or how to use it effectively, [here's an extensive guide about it](https://huggingface.co/hollowstrawberry/stable-diffusion-guide/blob/main/README.md#index). |
|
|
|
Recommended additional resources: |
|
|
|
* [blessed2.vae.pt](https://huggingface.co/NoCrypt/blessed_vae/blob/main/blessed2.vae.pt) - without a VAE selected in the settings your colors will look faded. |
|
* [TagComplete](https://github.com/DominikDoom/a1111-sd-webui-tagcomplete) to view all available anime tags as you write your prompt. |
|
* [Image Browser (updated fork)](https://github.com/AlUlkesh/stable-diffusion-webui-images-browser) to view your past images and their prompt metadata. |
|
* [Lycoris extension](https://github.com/KohakuBlueleaf/a1111-sd-webui-locon) to be able to use locons and lohas, which are new types of lora. |
|
* [EasyNegative](https://huggingface.co/datasets/gsdf/EasyNegative/blob/main/EasyNegative.safetensors) - put it in your `stable-diffusion-webui/embeddings` folder and write `EasyNegative` in your **negative** prompt to drastically improve your images. |
|
|
|
I make Loras of my own, particularly multi-outfit or multi-character Loras. Many of the Loras found in this repo are made by me or my peers. [Here are some other Loras I made](https://huggingface.co/hollowstrawberry/multicharloras). |
|
|
|
|
|
|
|
# Models |
|
|
|
I merged these with hll4-p3-ep8, which is the newest vtuber-finetuned model. They have blessed2 VAE baked in. Many more mixes can be found [here](https://huggingface.co/grugger/chubas). |
|
|
|
* [HeavenOrangeVtubers_hll4_final](https://huggingface.co/hollowstrawberry/holotard/blob/main/models/HeavenOrangeVtubers_hll4_final.safetensors) |
|
* [AOM3_hll4_final](https://huggingface.co/hollowstrawberry/holotard/blob/main/models/AOM3_hll4_final.safetensors) |
|
* [AOM2hard_hll4_final](https://huggingface.co/hollowstrawberry/holotard/blob/main/models/AOM2hard_hll4_final.safetensors) |
|
* [Grapefruit4.1_hll4_final](https://huggingface.co/hollowstrawberry/holotard/blob/main/models/Grapefruit4.1_hll4_final.safetensors) |
|
|
|
<details> |
|
<summary>Click here for a comparison</summary> |
|
|
|
![A comparison of different vtuber checkpoints](comparison3.png) |
|
</details> |
|
|
|
|
|
|
|
# Loras |
|
|
|
Most useful 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:1>` |
|
Most Loras work well with a weight of `1.0`, but some older ones work best at `0.7`. |
|
|
|
Learn to make your own Loras [with my guide](https://civitai.com/models/22530). |
|
|
|
|
|
|
|
## Multi Outfit Loras |
|
|
|
These are the most modern and have several outfits of each talent in a single Lora, or the main outfit if no other options are available. Check the model page / text files for examples. |
|
|
|
### HoloEN |
|
|
|
* [Mori Calliope ×9](https://civitai.com/models/173344) |
|
* [Takanashi Kiara ×3](https://civitai.com/models/48195) |
|
* [Ninomae Ina'nis ×5](https://civitai.com/models/17922) |
|
* [Gawr Gura ×6](https://civitai.com/models/20447) |
|
* [Amelia Watson ×4](https://civitai.com/models/27398) |
|
* [Irys x4](https://civitai.com/models/111029) |
|
* [Ceres Fauna ×3](https://civitai.com/models/97377) |
|
* [Ouro Kronii ×3](https://civitai.com/models/124507) |
|
* [Nanashi Mumei ×4](https://civitai.com/models/124549) |
|
* [Hakos Baelz ×5](https://civitai.com/models/6080) |
|
* [Tsukumo Sana ×1](https://civitai.com/models/20175) |
|
* [Shiori Novella x1](https://civitai.com/models/128135) |
|
* [Koseki Bijou x1](https://civitai.com/models/129972) |
|
* [Nerissa Ravencroft x1](https://civitai.com/models/141707) |
|
* [Fuwawa Abyssgard x1](https://civitai.com/models/132928) |
|
* [Mococo Abyssgard x1](https://civitai.com/models/132419) |
|
|
|
### HoloID |
|
|
|
* [Ayunda Risu ×2](https://civitai.com/models/21209) |
|
* [Moona Hoshinova ×1](https://civitai.com/models/124535) |
|
* [Airani Iofifteen ×1](https://civitai.com/models/27558) |
|
* [Kureiji Ollie ×4](https://civitai.com/models/28686) |
|
* [Anya Melfissa ×3](https://civitai.com/models/27935) |
|
* [Pavolia Reine ×3](https://civitai.com/models/15981) |
|
* [Vestia Zeta ×3](https://civitai.com/models/52609) |
|
* [Kaela Kovalskia ×1](https://civitai.com/models/28355) |
|
* [Kobo Kanaeru ×3](https://civitai.com/models/145897) |
|
|
|
### HoloJP |
|
|
|
* [Tokino Sora ×3](https://civitai.com/models/19432) |
|
* [Roboco-san ×4](https://civitai.com/models/154115) |
|
* [Sakura Miko ×4](https://civitai.com/models/33471) |
|
* [Hoshimachi Suisei ×8](https://civitai.com/models/11765) |
|
* [Azki ×2](https://civitai.com/models/26419) |
|
* [Yozora Mel ×3](https://civitai.com/models/21431) |
|
* [Shirakami Fubuki ×9](https://civitai.com/models/156406) |
|
* [Natsuiro Matsuri ×5](https://civitai.com/models/23883?modelVersionId=28543) |
|
* [Akai Haato ×5](https://civitai.com/models/6489) |
|
* [Aki Rosenthal ×2](https://civitai.com/models/124521) |
|
* [Minato Aqua ×11](https://civitai.com/models/17816) |
|
* [Murasaki Shion ×4](https://civitai.com/models/105212) |
|
* [Nakiri Ayame ×6](https://civitai.com/models/12658) |
|
* [Yuzuki Choco ×1](https://civitai.com/models/20305) |
|
* [Oozora Subaru ×8](https://civitai.com/models/22332) |
|
* [Ookami Mio ×6](https://civitai.com/models/156824) |
|
* [Nekomata Okayu ×6](https://civitai.com/models/23962) |
|
* [Inugami Korone ×7](https://civitai.com/models/153356) |
|
* [Usada Pekora ×9](https://civitai.com/models/27247) |
|
* [Shiranui Flare ×2](https://civitai.com/models/20509) |
|
* [Shirogane Noel ×6](https://civitai.com/models/114191) |
|
* [Houshou Marine ×8](https://civitai.com/models/47510) |
|
* [Uruha Rushia ×9](https://civitai.com/models/36097) |
|
* [Amane Kanata ×7](https://civitai.com/models/124532) |
|
* [Tsunomaki Watame ×5](https://civitai.com/models/9430) |
|
* [Tokoyami Towa ×6](https://civitai.com/models/156139) |
|
* [Himemori Luna ×2](https://civitai.com/models/124564) |
|
* [Kiryu Coco ×5](https://civitai.com/models/97114) |
|
* [Yukihana Lamy ×4](https://civitai.com/models/16876) |
|
* [Momosuzu Nene ×4](https://civitai.com/models/40437) |
|
* [Shishiro Botan ×4](https://civitai.com/models/10390) |
|
* [Omaru Polka ×3](https://civitai.com/models/124570) |
|
* [La+ Darknesss ×1](https://civitai.com/models/14019) |
|
* [Takane Lui ×1](https://civitai.com/models/17673) |
|
* [Hakui Koyori ×3](https://civitai.com/models/21233) |
|
* [Sakamata Chloe ×4](https://civitai.com/models/124502) |
|
* [Kazama Iroha ×3](https://civitai.com/models/149755) |
|
|
|
### Others |
|
|
|
* [A-chan](https://civitai.com/models/27927) |
|
* [Harusaki Nodoka](https://civitai.com/models/33130) |
|
* [Mano Aloe](https://civitai.com/models/35099) |
|
|
|
|
|
|
|
## Small Loras |
|
|
|
Some older Loras were manually 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. |
|
|
|
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. |
|
|