This model was trained on screenshots from Botw and a few from Hyrule warrior.
Big thanks to Mety333 for gathering quality training images.
You can check out his stuff on twitch or youtube.
You can pick either V1, which is easier to use with less prompting, but the results will not be that diverse in composition (mostly upperbody only), and tends to look a bit blurry.
Or V2, which is hardest to use and require more prompt tuning, but the results are clearer, with more variety in composition, and it can just do more in general.
To reference the art style, use the token: botw style
You can check out the diffuser model here on huggingface