This model is designed primarily for creating covers and visuals for music videos in the "Haru" style - essentially for the visual presentation of music.

It was a challenging model to create; it took six attempts to get it right. This LoRA doesn't function quite as flawlessly as a standard character LoRA, as it incorporates a specific style and overall concept alongside the character itself -and the character is quite complex for an SD model. The LoRA is optimized for a 16:9 horizontal aspect ratio; while it can work in a vertical format, it tends to produce more errors.