This is the LyCORIS version of a previously released <a target="_blank" rel="ugc" href="https://civarchive.com/models/66944/blacksouls">LoRA</a>. This one came out a bit different, besides being a bit more versatile, the style is a somewhat milder and the scribbles of the original images seems to have been turned into extra details instead. Pick this or the LoRA depending on what outputs you're after.How to use Just prompt characters as normal. There's a lot of various monster and animal girls in the dataset so those will work fine. There are various ways to change output, Euler A (gives a more smooth look), DPM++ 2M Karras (more drawing-like look) and finally adding "3D" and/or "realistic" in negatives to push the images further towards the original style. As usual, how well things work is in the end up to how well your model handles drawing styles. For further prompting tips see the sample images or uploaded zip of training tags.MISC Trained on NAI at 512 resolution with a dataset of 192 images. All images are untouched txt2img, no inpainting or helper embeddings used for quality transparence, e.g. - things should look even better if you do. The training was done on a small subset of the artists images, as such the output of this LoRA is not an accurate representation neither in quality or style of the source material. Go support the <a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/2938678/artworks">artist</a> if you want the good stuff.

Files with SHA256: 336fd11c9e9868b0c3f5e44f0b80166958525e9af8479ce916f1a71de50f2b3d