Additional model for better results with Joey, keywords is also "joey". Got best results when using both LoRas with 0.65 weight
xxdd_ignore_joey.safetensors