Preview images don't use any other LoRAs.<h1 id="commercial-use-of-this-model-is-not-allowed">❌💲 Commercial use of this model is not allowed</h1>I ran a prompting test: "fearinsideout" alone is not enough to get accurate appearance. His floating eyebrows, long nose and slim body seem to work without being prompted for. If you don't describe his clothes as they are in canon and his eye colour they WILL change. If you don't specify his skin colour it may change as well. I also got random ears with the shorter experimental prompt.Fear might be my favourite character in the Inside Out franchise, but I noticed there wasn't a single model of him, despite the franchise now having 2 feature films and a short. At first I thought it may be because he's not seen as conventionally attractive or cute, but I found 3 models of Anger, so I guess it's just that nobody cared. 🤷‍♀️<h2 id="dataset">🖼 DATASET</h2><ul><li>55 pictures total from both feature films, which are property of Disney and Pixar. Tagged manually.</li><li>Dataset does NOT include alternate outfits (it's only Fear in Riley's mind in his regular outfit). I think about making an updated version with more outfits.</li><li>No editing besides basic cropping.</li><li>Gathering fullbody solo pictures (without them getting repetitive) turned out a lot more difficult than I expected, so there's a decent chunk of pictures where he's at the control panel, and in some pictures parts of other characters (especially Anger) get in.</li><li>Aside from the trigger word I used these captions to describe the character: "anthro, male, slim, lavender skin, purple eyes, long nose, floating eyebrows, antenna, grey pinstripe shirt, black and white houndstooth vest, magenta bowtie, dark blue pants, black shoes". His antenna changes shape, and even when it's spiral (most common shape) the direction is inconsistent, so I have low expectations with it.</li></ul>

Files with SHA256: cc9e5c01abf384e173ba131ed4c3559327ba16d663bf41d26bcf5aefe1790465