That model was trained on 13 pictures using AOM3
koboKanaeruVtuber_kobo.safetensors
Kobo-000015.safetensors