This LoRA was trained on about 65 images of the Lodoss OVA. Locally 2400 steps, pruned from 64 dimensions to 32.Trigger word: l0d0ssw4rCharacters (that work well)<ul><li>Deedlit</li><li>Parn</li></ul>If you want the styling to be even more "classic" you can use additional triggers like<ul><li>retro artstyle</li><li>1990s (style)</li></ul>

Lodoss OVA Style (Anima)

Trained with more images for a slight longer time. This should result in a more consistent style experience.

Files with SHA256: e3f1f43ec666e55862d1d127f65209e7fe35ed268bbc25fe9970cc4cff68f107