I tried to draw a character recently and had this idea to generate reference poses! This is an experimental LoRA.
This LoRA was trained almost exclusively on the female figure, and was intended to be used standalone. honestly not sure how it does with other LoRAs. Adding wireframe to the prompt adds the drawn lines across the body, but sometimes the lines are mistaken for hair or part of the head haha.
It's not the most consistent, but it gives me the references I want. Figured I'd share it.
EDIT (1/14/25): There is a noticable watermark that can appear when generating images using this LoRA without using quality enhancing tags.
Illustrious XL:
Prompts:
Trigger word: r3ference
Recommended: 1girl, wireframe, white background, faceless female
Info:
Trained on illustriousXL_v01 using 87 images, 15 epochs, and 1 repeat (1305 steps total).
Strength: Works very well with 1.0.
Description
FAQ
Comments (4)
this is amazing!
It has some watermark from "AnimeReferencePossescom" if people dont mind.
Yes I did forget to address that! I only saw the watermark on generations without my quality tags applied.
Great Lora. I'm sure this will be helpful as reference when i'm drawing something.
I sometimes get too distracted when referencing other artwork but this helps get an idea of the pose without having any distractions in the reference picture.
The guidelines from the wireframe are also surpisingly correct which is great to help with depth perception.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.