<h1 id="commercial-use-of-this-model-is-not-allowed-g0zk5wrs2">❌💲 Commercial use of this model is not allowed</h1>Preview images use different checkpoints, but no other LoRAs.<h3 id="warnings-uyip7bw5t">⚠WARNINGS</h3><ul><li>Common issues: human face features, arms, bad or missing mouths, wrong number of eyebrows or pupils (or missing entirely). Unfortunately these LoRAs are kinda weak, I found myself increasing their weight often.</li><li>The SDXL version has a high risk of giving double tails.</li><li>I thought this could be a higher quality counterpart to the <a target="_blank" rel="ugc" href="https://civarchive.com/models/599407/worms-team17-mayhem-pdxl">Mayhem LoRA</a>, but it's much more difficult to work with. :(</li></ul><h3 id="dataset-qj1su5ovl">🖼DATASET</h3>Trained on images of Worms renders (all captioned), mainly those related to:<ul><li>Worms 3D</li><li>Worms Forts: Under Siege</li><li>Worms: Revolution / Clan Wars / Battlegrounds (G4 games where worms come in different shapes)</li></ul>... so the way worms look in the dataset images have some variety. Some pictures have 2 worms, there's also one with 5.❌ Dataset does NOT include:<ul><li>G1 material</li><li>G2 material</li><li>Worms Rumble material</li><li>any in-game images</li><li>any fanmade models (all images are property of Team17, with minor edits and upscaling)</li></ul>

Worms Team17 render mix [SDXL-PDXL-IlluXL-NoobXL]

53 images. The dataset images in general are lower quality. They also contain Heavy and Scientist worms, whom I removed in the retrained version.This is actually the 2nd training attempt. This was before I started generating most images locally, so when I saw the sample images I thought they were trustworthy (they have NEVER been) and assumed the LoRA was a failure without actually trying it out. So my 1st training attempt just went to waste.Trained against SDXL.

Files with SHA256: c173af6bb41ad8df1ac7bd12241cb924b718244c0bff7c879b80af269b2a8f57