LARP (Low-rank Adaptation for Realistic Photography)

This was a Krea 2 Hillclimb for ultimate realism and detail

• Trained to work exceptionally well when combined with twotensors.ai's forensic cleaner when using for social media

• Extremely high rank of 384 for maximum detail extraction (can also backfire for ~10% of prompts)

• 16k steps at BS=2 on ~6000 images captioned with GPT 5.6 Terra + other trainer modifications

• Trained on RAW, generated on Turbo

Generation Tips:

• This model is NOT for the GPU poor. Generations look significantly better the higher resolution you generate at (~3k optimal, it falls apart after that)

• Stay at 8 steps, going over 8 steps can add TOO MUCH detail.

• Prompts must be EXTREMELY detailed, most of the prompts in the post are from https://www.promptlibrary.space/

• Untested on Image editing workflows