LoRA trained on some classic expressions.It's not perfect but it's the best I could get. I have attached my training data, if anyone thinks they can perfect it, they're welcome to try.Backgrounds are a bit finicky. Sometimes, it zooms out very far for no reason when you prompt for a background, especially on lower resolutions.These are the expressions (and also the trigger words):&gt;_&lt;@_@=_= (Combine with "closed eyes")X X (Space, no underscore):Ɪ (That's an "i". Combine with "pout")._. (Works somewhat; Chibi only)JitomeWeight ~1.Trained on AnythingV3.

Files with SHA256: b1126cbc742a94f0c09c2db7232d37b7d2e0c10b62b54f044319b09cf6bc0f9b