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: b3e05a2da84c1aff182675f48bfcbc2f03108e04837f95e9febce08c7a270f18