This is a Flux dev Lora trained on the film 'Loving Vincent'.

Use this Lora at Guidance: 2 for maximum strength

You can choose to include the trigger word <vncnt> in your prompts, but it is not necessary. The style will show regardless in most cases. I only use the token when the subject I'm generating is overpowering the style entirely.

This Lora is very flexible and has taken no hit to hands, complexity, or quality.
The model was trained ONLY on carefully curated screenshots from the film.