In my pursuit of training a new Furry model I've begun to produce new loRA's, in preparation for creating a more varied, cohesive database from which to work off of.
This model handles both Feral and Anthro variants rather well. Also, the tags might be a bit strong on this one, so you might want to turn down the strength of the LoRA depending on what you're going for. Using the weights at 0.4 seems to work best for more varied results.
This LoRA was trained using Furry Vixens 2.0, and works exceptionally well with it! The images were generated using clip skip 2.
Image resolution for training was 512x786
You can use these tags when generating images:
Eevee, Eevee tail, Eevee ears, brown eyes, brown fur, autumn, forest, trees
Though obviously all other tags work well too! These might get you some more accurate Eevees though, depending on the rest of your prompt.
And be sure to use these settings!
Steps: 32
Sampler: DPM++ 2M Karras
CFG scale: 7
Size: 384x512
Model: Furry Vixens v2
Denoising strength: 0.6
Clip skip: 2
Hires upscale: 2
Hires upscaler: R-ESRGAN 4x+ Anime6B
Positive Prompt
Eevee, Eevee tail, Eevee ears, pawpads, smile, open mouth, sitting, looking at viewer, autumn, forest, trees, <lora:Eevee V1:0.7>Negative prompt
(nsfw:1.4), (worst quality:1.4), (low quality:1.4)
Initial release.
Trained using Furry Vixens 2.0.
1. The rights to reposted models belong to original creators.
2. Original creators should contact SeaArt.AI staff through official channels to claim their models. We are committed to protecting every creator's rights. Click to Claim
