A model of Misa Campo, one of my long-time favorite models. You should use this LoRA with the <a rel="ugc" href="https://civarchive.com/models/25694/epicrealism">epicrealism </a>model, using this prompt structure for the positive prompt: <pre><code>Misa Campo, a sexy Asian woman, wearing &lt;something&gt; &lt;lora:misacampo:1&gt;</code></pre> This generally gives good results, along with the badhandv4, OverallDetail, BadDream, and UnrealisticDream negative embeddings. Below about 0.9, the face starts getting mangled, the anatomy goes wild, or both.

Misa Campo

Pruned the training dataset to make the face more consistent.

Files with SHA256: c2f6738ed4eb3f0125658b47afdeb0aa897f8b28c19c49264cbf99d3cf1390c9