The V1 Lora has been released. Please use V1 since it is more stable and more compatible with both 2d and 3d styles.
The V1 is trained on Stable Diffusion V1.5 with 32 epochs with about 60 images.
To trigger specifically the exact outfit of the character, use `thighhighs`, `bare shoulder` and `hair ornament` in the positive prompt.
Recommended weight for 3d style: 0.8-1.0
Recommended weight for 2d style: 0.7-0.8
Recommended models: FaceBombMix-V2, MeinaMix, AOM3, Guofeng3-V2
<=====================================>
NSFW should work as well but not be included in the sample images. You may try it your self :)
<=========== V0.1 info as below ==============>
This is the Lora trained on AOM3 model with 14 epochs for the character Cai'er in the animation "Throne Of Seal".
Recommended to use it with model AOM3 or CounterfeitV2.5 with weight 0.6 - 0.8 to maintain the original dress.
If you would like to use it with another Lora or with a different dress style, a weight of 0.4 - 0.6 could be better.
To make the style more 2d, just add "3d, realistic" in the negative prompt.
This is model is created for personal interest in learning Lora or animation fan art, not for commercial purpose.
Description
The first version for the one trained on AOM3. Mostly stable but maybe overfitting for large weight. Decrease the weight of the Lora gradually to find the best weight settings for your model.
FAQ
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.