This LoRA model is quite well-trained, though there are some mistakes in the generated artwork.
For instance, the generated image might include a drop of sweat on the character's face. To address this, you can add "(no sweat drop)" to the prompt or include "(sweat drop)" in the negative prompts.
Similarly, it often generates glasses on the character, whether specified in the prompt or not. To prevent this, you can include "(no glasses)" in the prompt or add "(glasses)" to the negative prompts.
Hope this helps!
