Removes (positive side) or Add (negative side) censorship from genitals.Works nearly always on mosaic censoring.Might need to add a bit more bar censorship examples to the dataset so it performs better on those.Placeholders replaced during training in captions:<pre><code>[slider]
placeholder = "&lt;CENSOR_AMOUNT&gt;"
positive = "exposed, uncensored"
negative = "mosaic censoring, bar censor, censored"
neutral = ""</code></pre>

Files with SHA256: 593decfdd9c57477669b716ed51858a7a2d212f810ec2c3288358c86a98936fb