Emily Bloom - MetArt

Trigger: em1lyblm

LoRA Strength: 0.80 - 0.90 (sweet spot seems to be around 0.8 to 0.85)

LoRA model was trained on CyberRealistic Pony v8. I decided to no longer train off of the Pony Base model and have now moved to using the CyberRealistic Pony model and the results have improved significantly. The LoRA has much better realism and resemblance.

You may need to use a FaceDetailer for better results.

Version 1:

This LoRA is SPOT ON. It's almost perfect resemblance. Tagged everything but eyes and hair color. Kept those to improve resemblance.

Dataset images are from her modeling years from 2014 to 2017.