A character LoRA trained for cinematic portrait and video generation. Compatible with HunyuanVideo / FreeVid. Produces consistent identity with strong facial detail and natural lighting. Recommended strength: 0.8–1.2.
Description
Initial release of the nanabanana_xl LoRA trained for HunyuanVideo / FreeVid. Provides consistent identity and cinematic detail. Recommended strength 0.8–1.2.