I'm hoping with this lora to achieve the feel of Anhes' bara art without sacrificing flexibility and ease of use. This version limits the dataset only his soft shaded bara art. So no Lush Puppies, digimon, etc. Lush...
Model source
Source description
Uses the same dataset as the Anima version, so it prefers tag based prompting. I recommend using at 0.8 weight for most cases / models. There's also an annoying black box that can rarely appear on the top left. I trained the model twice to get rid of it and it still appears sometimes. Sorry about that.
Sources
1 sourceModel artifacts
1 artifactAnhes_Krea2_V10.safetensors
safetensors · BF16 · 109 MB · SHA-256 f2c9653c2035…ffcf · Civitai
DownloadSource excerpts
1 excerptSource context: 1523 downloads · Type LORA · Base model Krea 2