<h3 id="heading-2">Updated</h3><p>Used PonyXL to train this one. Updated dataset to achieve better performance. It enhanced this concept on PonyXL and related models. See the examples first.</p><p>I should say the Pony model fixed many problems I met. I do advice people to give it a try.</p><p></p><h3 id="heading-634">Introduction</h3><p>An experimental LyCORIS to generate futanari filling condom. Use the key words to control the size of penis and condom.</p><p>Still a lot of room for improvement.</p><p></p><h3 id="heading-230">Pros</h3><ul><li><p>Can be combined with multiple Loras</p></li><li><p>Little impact on the art style</p></li></ul><h3 id="heading-97">Cons</h3><ul><li><p>Using 3D materials for training, the art style may become strange under high weights</p></li><li><p>Twisted penises and multiple penises</p></li><li><p>Bad glan shape</p></li><li><p>Lots of trials needed</p></li></ul><p></p><h3 id="heading-635">How to use it</h3><p>I recommand using it in weight 1-1.2.</p><p></p><h3 id="heading-636">Training details</h3><p>Traingset is of about 107 images. Half of them are chosen as regularization images. Total steps are about 7600.</p><p>resolution=768*768</p><p>batch_size=2</p><p>epoch=10</p><p>network_dim=32</p><p>network_alpha=16</p><p>conv_dim=8</p><p>conv_alpha=4</p><p>clip_skip=2</p><p>Using Prodigy as optimizer, setting as the author recommend.</p>

Files with SHA256: 78547f7f7e5546d55310b79e614f4d09ed7af14d1736169ca2ffff839b065f3e