This is a LoRA that adds photo/realism to the model you use it with.

The intended use is to add some 3D and realism to anime pony models like Autismmix.

It works with vanilla Pony v6, although the results are not amazing as vanilla Pony is too raw.

Recommended Settings

For most cases, I recommend using 0.5 in weight.

The higher the weight, the more 3D, realistic, and usually zoomed in the result will be.

There is no trigger word. Just apply the LoRA. <lora:PoenyPDXL:0.5>

Training

The model was trained with a dataset of long, detailed, complex captions using a custom instruction to a Claude model, using Poe. This gave it the name of Poeny.