A simple LoRA model trained on the character portraits for Lost Ruins. The dataset is only 53 images so replicating the style is a bit of a challenge. I also trained in all the character names and Lost Ruins but I find they don't improve results much.Trained using AnythingV3 as a base. Sprites were collected from <a target="_blank" rel="ugc" href="https://www.spriters-resource.com/pc_computer/lostruins/">The Spriters Resource</a>.

Files with SHA256: 5bcb50d9b67831651ec8c82df39e12c7283865060573288597bed6ef875861a0