The i5-xl large model is based on NTR MIXxi and NNAillous-xl.Image Distribution:Realistic: 22%Cartoon: 22%Composite Images: 50%Objects and High Resolution: 6%Due to the high proportion of composite images, it prevents realistic images from being contaminated into a plastic-like cartoon style. After training with the i5-xl LoRA model, it can achieve separation and distinction between characters, objects, and backgrounds, thereby creating a clear contrast between realism and anime.Stability level: approximately 65%Model Essence:Main Model: _Illustrious-xlParent Model: _NTR MIXxi and NNAillous-xlChild Model: _i5-xlRecommended Prompt Order: parametersquality, person count, character names, background, tag Negative promptLow quality and resolution, etc.Lazy Prompt Suggestions:Make the most use of a standard AI chatbot and ask it to become a Danbooru prompt master. Then, the content of the prompts will have been checked for correctness and objectivity.Sampling Suggestions:Euler aDPM++ 2M SDESteps: The more steps, the more sharpening you'll get.CFG Scale: The larger the scale, the higher the contrast will be.Tool Recommendations:ComfyuiCustomization: HighConfiguration Demand: LowStable Diffusion A1111Customization: LowConfiguration Demand: HighThese are subjective recommendations; over time, many innovative tools will emerge for you to consider and use.
