Licious Wickerbeast

Ever wanted to wicker it up, but couldn't quite get it looking right? This may just be the ticket for you! Designed for the Illustrious model from the outset!

Trained off 142 3d images of various wickerbeast models, this LoRA was designed to focus on the key features of the wickerbeast! To trigger it, you can just type in wickerbeast in your prompt; if you want to refine it to only male or only female, just use the tags wickerlord and/or wickerlady!

Note for v2A: The trigger words lusc_wicker, wickerlord and wickerlady were removed as superfluous - to trigger the wickerbeast features, use the terms wickerbeast, clawfingers, jagged maw and wicker tail instead!

Caveats:

• You may find it tricky to change the hair, as all of the reference images were derived from a specific wickerbeast model set, with a specific hair style on almost all variants. If the hair continues to be problematic, you may wish to use in-painting, masking and image editing to get the hair style you want onto your wickerbeast. Fixed in v2A! Now the wickerbeast can have just about any hairstyle you can imagine! Don't forget that you can bracket your selections to amplify them; for instance, you can make a prompt like (absurdly long hair, hime cut, blue hair:1.5, gold hair:0.5,) to create a very long hair style with a hime cut, and a mix of blue and gold hair with more blue than gold.

• The original fur patterns of the training images may appear if specific fur colors and patterns are not specified in your prompt; be sure to be specific about your fur patterns where possible!

• There is also a tendency for some checkpoints to add an animal nose (canine specifically) to the snout; if this occurs, add animal nose, canine nose or colored nose to your negative prompt, adding emphasis to them as needed to make certain.

• Because all of the training images are 3D images, this will bias output images to be 3D as well. To prevent this from happening, add the 3D tag to your negative prompt Changed in v2A: 2d images have now been added to the training data to help vary the appearance! However, adding the 3D tag to your negative prompts will still be useful for adjusting your output's appearance.

• Additionally, many of the training images were NSFW in nature, so if this is undesired, be sure to add the SFW tag to your positive prompt, along with clothing as needed. You can also add the NSFW tag to your negative prompt as an added insurance.

• This LoRA was trained specifically for the Illustrious and Pony base models; please ensure you pick the correct LoRA for your specific checkpoint, as they are not guaranteed to work between base models! The unique trigger tags are the same between all versions of the Licious Wickerbeast LoRA!

Thank you for having a look! If you generate something fun with this LoRA, whether here on CivitAI or externally via local generation, please share them here - I'd love to see what you get the generators to show you!