RealCosplay is multi-style base models, not limited to realism alone.
I don't want to destroy the anime concept of illust or noob,like other Lora or Merge model. So i finetune this model in anime style by default.I'm aimming to use high-quality real photo data to achieve realistic style, and improve anime concept's anatomical\scene composition, through its realistic concept.
Recommended settings:
• postive prompts for realistic
realistic ,cosplay,
• negative prompts
• (for realistic) illustration,anime,sketch,comic,cartoon,toon,oil painting \(medium\),flat color,outline,3D,2.5D,2D,unreal,unity,blender,anime,smooth skin,semi-anthro,impasto,
• (optional)ai-generated,abstract background,worst quality,normal quality,lowres,jpeg artifacts,blurry,fading,mature,
• Suggest parameters
Euler a
cfg : 4~5
Version RealCosplayChenkinV1 2025-12-06
• Recommend using this version – it’s a huge improvement over the old one.
• Full finetuning from Chenkin Noob XL V0.1
• The realistic trigger prompt is "realistic" and "cosplay". When generating realistic styles, Please include the recommended negative prompts. Because this is a realistic enhancement base model, not a realistic replacement base model.
• Expanded Training Data (now 11,666 images 60% realistic \ 30% anime \ 5% 3d \ 5% other)
• Caption with tag+nl (pixai-tagger-v0.9 + Huihui-Qwen3-VL-30B-A3B-Instruct-abliterated Q8), Manually modified 50% of the captions ,still updating.
Version RealCosplay1.2 2025-06-03
• Full finetuning from noob eps 1.1
• The trigger prompt has been changed to "photograph". The previous "realistic" is closer to semi-realistic, which will interfere with the photo-realistic style.
• Compared with version 1.0, each has its own advantages. Just use the version you like.
• The following rare concept prompt words have been trained
Brightness:darkness,dark,dim,bright,overexposure
Background:blurry background,depth of field,bokeh,faded background,detailed background,fog
Composition:fisheye,partia