Version 0.1 (Early / Experimental Release)
A character LoRA for Link, based on his appearance in the upcoming The Legend of Zelda: Ocarina of Time remake. This is a very early release. The current dataset is still limited, and this is also the first Krea 2 LoRA I’ve trained, so a lot is still in flux. For now, this is what I have - and I hope people enjoy experimenting with it! Feedback comments welcome.

Recommended Settings
Trigger: L012
LoRA strength: 0.6 - 1.0
Prompt: See examples uploaded.

About v0.1
This should be treated as an experimental first version, rather than a finished model. The dataset is currently fairly small, so you may see inconsistencies depending on:

- Pose and camera angle.
- Facial expression.
- Lighting.
- Clothing and equipment
- More unusual compositions.
- Scenes far outside the current training data.

That said, the model is already capable of reproducing Link across a reasonable variety of poses, expressions, perspectives and environments.

Future Versions
More versions are planned. I want to expand the dataset and improve:

- Facial consistency.
- Likeness from difficult angles.
- Clothing and equipment details.
- Hands and small details.
- Side and rear views.
- Action poses.
- Expression variety.
- Consistency across different lighting.
- Flexibility with alternate clothing and environments.

Because this is my first Krea 2 character LoRA, I’ll also be adjusting the training process itself as I learn more about what works well with the model.