Warning: This model is a bit difficult to work with.

I created this because I wanted to know what the minimum amount of art a LoRA can be trained with, and I'm not really sure if this is it or a bit under it. This is a LoRA for the character Linda from the GameCube game "Custom Robo: Battle Revolution" or just "Custom Robo" in the west.

This LoRA was trained on one official art of her, two low quality in-game portraits, one nsfw fan art by kreschun, one screenshot from the game, and a few 3d renders I made using her in-game model. That's basically the full extent of what can be used for training data for her.

Trained on nai, example images made with Anything v3, AOM2-nutmegmixGav2+ElysV3, and corneos7thHeavenMix