!!! Please do not use this LORA for commercial purposes !!!
This model leans toward a realistic style, which can also be influenced by other style-based LoRAs.
You may refer to the prompts used in the attached images for best results.
With the release of the Z-image Controlnet , this has now become the most efficient and highest-quality portrait model available.
After extensive training and tuning, it takes 1500 steps to produce a high-quality LoRA. In the future, I plan to train both Illustrious and Z-image versions.
I hope you enjoy using it and appreciate your support!