This LoRA responds particularly well to the "realistic" tag, if you want less flat anime faces, that is the key. This is particularly useful when character tags push the style towards a more flattened anime aesthetic. Trained on Anima Preview 3. I do not recommend using RDBT with this LoRA, it flattens the effect way too much for this type of style. All the images have metadata, and include a workflow you can use that has a very useful highres fix pass that is stable and brings out lots of detail. You don't need any crazy workflow, it works just fine with the default ComfyUI workflow, or any other tool you have. <h2 id="trigger:-@45ur4u53r">Trigger: @45UR4U53R</h2>

Files with SHA256: 4e4eadc909dc5c00659037adbd9f8d852e880aab4c52e8e77d7cb37a27dd2060