Help me grow on this platform by leaving a like and following me 👍.
Addendum (17/01/2025): Actually the visually accurate version work just fine, i was using a weird forge webui back in September of last year. Came back to auto 1111 and it work just fine even with photorealistic models like photon without this weird paint effect every time, i've update the showcase gallery with some new pics 👍 .
About the visually accurate version
This Version is way more accurate to the video game Shelob, but because of that it lack versatility and is unstable on some model.
It was trained on Civitai with 401 pics of the character Shelob with a resolution of 768*768 to get her Facial features right.
The trigger is Shelob woman:
And to get the exact clothing style of shelob you can add the prompt:
"black fringe robe with long pending arm laces shrug".
It is advisable to reduce the strength of this lora to at least 0.7 if you want any hope of photorealism and/or adding sunshine and colorful stuff to your generation. And to use adetailler for the face.
If you need help to see how to prompt, please refer to the pictures I have provided to illustrate the Lora. As always my Loras metadatas are never hidden.