This is my first proper attempt at training a LoRA using Civitai's LoRA training tool.

A relatively high weight is recommended: 0.8 - 1.0. No trigger words needed.

It doesn't do nearly as good a job of replicating her actual legs as the embedding I produced many months ago, but it does a much better job of avoiding steering your image toward lycra, and spiderman colours and patterns, affects other parts of the image (such as the face) less. The two differences may well be for the same reason, i.e. it doesn't have a strong influence generally.

Feedback very much welcome!