NOTE: click the preview images for better resolution, they are amazingly sharp (none of the preview images use upscaling or other post processing).
This is the first public version of my photorealistic model. It is a combination of merging and fine-tuning, with a strong influence from merging in Juggernaut XL v9, a model with unprecedented detail. It's optimized for photographs and able to represent some really unique concepts. It has parts of over 50 models recursively merged in, along with added training and custom LoRAs.
Steps: ~25-40
Sampler: DPM++ 2M SDE Karras
CFG: 3-11 (surprisingly coherent at higher CFGs up to 17)
Hi-Res - not necessary
ADetailer - useful for far away faces
