anime

NOTICE: My LoRAs require high amount of tags to look good, I will fix this later on and update all of my LoRAs if everything works out.

General Information

Overview

Welcome to the place where I host my LoRAs. In short, LoRA is just a checkpoint trained on specific artstyle/subject that you load into your WebUI, that can be used with other models.
Although you can use it with any model, the effects of LoRA will vary between them. Most of the previews use models that come from WarriorMama777 .
For more information about them, you can visit the original LoRA repository: https://github.com/cloneofsimo/lora
Every images posted here, or on the other sites have metadata in them that you can use in PNG Info tab in your WebUI to get access to the prompt of the image.
Everything I do here is for free of charge!
I don't guarantee that my LoRAs will give you good results, if you think they are bad, don't use them.

Installation

To use them in your WebUI, please install the extension linked under, following the installation guide:
https://github.com/kohya-ss/sd-webui-additional-networks#installation

Usage

All of my LoRAs are to be used with their original danbooru tag. For example:

asuna \(blue archive\)

My LoRAs will have sufixes that will tell you how much they were trained. Either by using words like "soft" and "hard",
where soft stands for lower amount of training and hard for higher amount of training.

More trained LoRA is harder to modify but provides higher consistency in details and original outfits,
while lower trained one will be more flexible, but may get details wrong.

All the LoRAs that aren't marked with PRUNED require tagging everything about the character to get the likness of it. You have to tag every part of the character like: eyes,hair,breasts,accessories,special features,etc...

In theory, this should allow LoRAs to be more flexible, but it requires to prompt those things always, because character tag doesn't have those features baked into it. From 1/16 I will test releasing pruned versions which will not require those prompting those things.

The usage of them is also explained in this guide:
https://github.com/kohya-ss/sd-webui-additional-networks#how-to-use

SocialMedia

Here are some places where you can find my other stuff that I post, or if you feel like buying me a coffee:
Twitter
Pixiv
Buymeacoffee

Plans for the future

LoRAs

Genshin Impact

<img src="https://huggingface.co/YoungMasterFromSect/Trauter_LoRAs/resolve/main/LoRA/Previews/zho.png" width="512" height="768"> <details> <summary>Sample Prompt</summary> <pre> masterpiece, best quality, zhongli (genshin impact), solo, 1boy, bangs, jewelry, tassel, earrings, ponytail, low ponytail, gloves, necktie, jacket, shirt, formal, petals, suit, makeup, eyeliner, eyeshadow, male focus, long hair, brown hair, multicolored hair, long sleeves, tassel earrings, single earring, collared shirt, hair between eyes, black gloves, closed mouth, yellow eyes, gradient hair, orange hair, simple background Negative prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts,signature, watermark, username, blurry, artist name, worst quality, low quality, extra digits, loli, loli face Steps: 22, Sampler: DPM++ SDE Karras, CFG scale: 7, Seed: 88418604, Size: 512x768, Model hash: a87fd7da, Denoising strength: 0.58, Clip skip: 2, ENSD: 31337, Hires upscale: 2, Hires upscaler: Latent (nearest-exact) </pre> </details>

- [Examples](https://www.flickr.com/photos/197461145@N04/albums/72177720305311423)
- [Download](https://huggingface.co/YoungMasterFromSect/Trauter_LoRAs/tree/main/LoRA/Genshin-Impact/Zhongli)

Blue Archive

Fate Grand Order

Misc. Characters

Artstyles