This is an SDXL version for Super Long Hair LORA, intended to replace existing LORA that was built under obsolete SD1.5. Trained with Civitai LORA training service, uses SDXL base model for training, everything in default settings and pre-configured with according datasets and its models. Should be work with most SDXL models, however, its strength settings must be set at 0.3 value. Clip skip set to either 1 or 2 depending what kind of models either anime or realistic. Resolution should be at least 1024x1024 or higher. Recommended to use this LORA with Civitai own image generator (strength should be default at 1.0).Also, please enable these settings on Auto1111's SD WebUI for best performance (both are found on Settings>Optimizations section):FP8 Weight (Drastically reducing VRAM usage, allowing GPU cards with 6GB VRAM can be used.Cache FP16 Weight for LORA
