Trained with 28 images, using just the standalone keyword is sometimes enough, but using more specific descriptions yields better results. The writing works pretty well with Anima, just write something like "the shirt has "insert text here" written on it".Anima prompts:<pre><code>hololivecheerup, skirt, crop top,</code></pre>Illustrious prompts:<pre><code>hololivecheerup, midriff, skirt, holding pom poms, </code></pre>

Files with SHA256: 4f568b7e603e6dbda403afcf5776989b7fc185495a3d05d89c4940d926bd3d11