These models are based on the audio training of Theresa's recording from Arknights and BiliBili
Base model