Apparently this LoRA works better with anime style models rather than photorealistic ones.
Beware of this when using it 😄.
Usually I use "wearing an eyepatch" in the prompt because more often than not the result is without it.
Tensor: https://tensor.art/models/634064978501967262