This is my first LoRa... so, please, be gentle with me.

This is Hei from Darker than black. This Lora works best in anime models, because it was trained with 0 realistic images. but it can do at least a decent job in realistic models.

In anime models the weight 0.6~ 1 works fine.

In realistic models, the ideal weight is 0.7 ~ 1, in 0.6 the overcoat just disappears and become a random piece of cloth.

Sadly, I can get his mask or knife to work yet. Why? Because I don't know how to use 2 or 3 activation words if most of the images don't include the subject of train. If someone could please teach me how to do that. So, I can try to create a new version of the model. If you have any other tips to help me to improve, please comment or contact me.

And also, You are free to improve this model or any other of my models I upload in the future. I can even provide my images with prompts if you need. The only thing I ask from you is... please contact me when it is ready. So, I can try it too. Hehehe.