Most of the time I prefer running this at 0.75 strength, in comfyui its much easier for me to test this but then I can't upload the images bc civitai can't parse the file.
Clothing is best not to use the prompts that I trained it with I found that creating a uniform for the waitress was better.
according to the LoRA file clip skip is zero not sure how this could have happened. I've been suspecting something is wrong with my json and might need to revisit this.
Known Issues: sometimes there will be an extra plate or tray near waist or above head I don't have a fix for that in sight. Very very rarely there might be text on bottom this more prevalent in V1. Sometimes there was a black artifact on the woman's breast on low cut shirts that is usually covered by hair.
Most of the time I prefer running this at 0.75 strength, in comfyui its much easier for me to test this but then I can't upload the images bc civitai can't parse the file.
Clothing is best not to use the prompts that I trained it with I found that creating a uniform for the waitress was better.
according to the LoRA file clip skip is zero not sure how this could have happened. I've been suspecting something is wrong with my json and might need to revisit this.
Known Issues: sometimes there will be an extra plate or tray near waist or above head I don't have a fix for that in sight. Very very rarely there might be text on bottom this more prevalent in V1. Sometimes there was a black artifact on the woman's breast on low cut shirts that is usually covered by hair.