<h1 id="heading-74">NOTE:</h1>v1.2 has been only tested using YiffyMix. <h1 id="heading-98"><a target="_blank" rel="ugc" href="https://tensor.art/models/666048582048179053">AVALIBLE ON TENSOR.ART</a></h1><h1 id="heading-1671">It's da pup gal!</h1>I wanted to LoRA-i-fy her for some time, it was a bit of a challenge since there's surprisingly not alot of fanart that could be used for Machine learning (or I just couldn't find much). Some detaset images also include custom fanart, so that's smthin.<h1 id="heading-1672">SOME INFO!:</h1><ul><li>It's hard to generate photorealistic stuff of her (for me at least)</li><li>doesn't work the best with cartoon models (for me at least)</li><li>Adding skye \(paw patrol\) into the prompt might help generate her correctly</li><li>main tag(s) are skyepawpatrol, goggles, pink vest, bottomless <s>(Bottomless tag probably isn't really needed when doing feral stuff)</s></li><li>Good loRA weight is 0.7-1.0</li></ul><h1 id="heading-1673">CONTACT INFO!</h1><a target="_blank" rel="ugc" href="https://twitter.com/CappyAdams">Twitter</a><a target="_blank" rel="ugc" href="https://www.deviantart.com/cappyadams">DeviantArt</a>

Skye // PawPatrol (LoRA)

Same detaset and everything as 1.1... however, this version was trained with YiffyMix v36. so the low quality issue should be gone

Files with SHA256: 2901c437fc724747f47a1a018a500243e87a939e2e0223e7cd2fdf465dad4852