Idea is to create a "vertical cut in" of characters at longer height generations (examples use 512 x 768). Maybe for a team attack? ;) Pretty straight forward. Activation word is vcutin but not 100% necessary. Recommend on anime models. Start with .80 weight and go from there. Can play with facial expression but complicated ones may not work. Example of things that work are devious smile, sly smile, straight face, angry, etc.Example Prompt: &lt;lora:vCUTIN-18:0.8&gt; a detailed character portrait vcutin, close up, detailed face, smug face&lt;lora:vCUTIN-18:0.8&gt; a detailed character portrait vcutin, detailed face, sly smileNEG: (blurry:1.3), monochrome, extra limbs, distorted, discolored, deformed, disfigured, watermark, (bad hands), (Worst quality, low quality:1.4)

vCutIn Concept

vCUTIN-18.safetensors

Got to do the horizontal cut version right? Made mainly because I like to burn my bread. Creates close up shot of characters for longer width generation (samples are 768x512). Anime models work best starting at weight at .80. hcutin key word but not necessary.Example Prompts:&lt;lora:HCUTIN-18:0.8&gt; a detailed character portrait hcutin, eyes close up, intense stare&lt;lora:HCUTIN-18:0.8&gt; a detailed character portrait, zoom in, eyes close up, intense stareNeg:(blurry:1.3), monochrome, extra limbs, distorted, discolored, deformed, disfigured, watermark, (bad hands), (Worst quality, low quality:1.4)

Files with SHA256: 85217fb8fb39c483d0caf593b03d7e15e086afe56614218a578ef6b17a20b145