Type | |
Stats | 110 1,275 |
Reviews | (11) |
Published | Dec 1, 2024 |
Base Model | |
Usage Tips | Clip Skip: 2 Strength: 0.4 |
Trigger Words | wolverine |
Hash | AutoV2 F5C20372FD |
First off, shout out to Freek22 for taking my data and training this LORA. This model is uploaded by permission. This LORA was trained on Ratatoskr 7.3 but it should work with any SDXL 1.0 based model. If you like this model, send tips to Freek22 for this work. Please credit Freek22 if using this LORA for any purpose.
This is a wolverine animal LoRA trained on natural animal photos. All the training data is safe for work photos of real animals. The goal was to help get the proper head, ear shape, face color and fur color instead of generations looking like unusual bears, raccoons, dogs or cats.
This LoRA is also useful to point the generations back in the proper direction for models that believe a wolverine is a black and yellow costumed human superhero with long metal claws jutting out of the back of the hands, or for models have no idea what a wolverine is supposed to look like in general.
Using the LoRA at full strength weight of 1.0 will push generations towards natural animals.
Use weights of 0.1 to 0.4 if trying to generate anthro, anthropomorphic, furry, were, or otherwise human-like wolverine hybrids. Depending on the model, this value can go higher to help with generation
Though the model was not trained on with any NSFW art, photos or illustrations, depending on the checkpoint used, NSFW content is still possible. Please prompt accordingly.