---
license: apache-2.0
---
This is the .safetensors version of Photomaker model. It is mainly used by stable-diffusion.cpp which can not read in the original .bin format.

Three tensor names are changed to better conform with the naming conventions used in SD models.

- vision_model.pre_layrnorm.bias    -> vision_model.pre_layernorm.bias
- vision_model.pre_layrnorm.weight  -> vision_model.pre_layernorm.weight
- visual_projection.weight          -> vision_model.visual_projection.weight