>>3085351All I said was that if a character LoRA is not faithfully recreating a character, there's an issue with the dataset or tagging. It's not a problem with the base model (Anima). Only a troll would disagree with this.
>>3085349I didn't contest this. I agree. Regardless, if a LoRA can't reliably reproduce a specific detail, it's an issue with the LoRA training, not the base model. But maybe you weren't contradicting me, and you just wanted to add a tangent to the chat? Idk.