<h3>LoRa trained on Nu-13 (ν-13) from Blazblue.</h3>Updated LoRa with a v2 now with more flexibility! With a pruned dataset and hand-picked captions for character training, it now eases the way it interacts with different clothing, colors, styles, and even most of the times with another LoRas. Triggers with Nu-13, red eyes, white hair, braid, ahoge, eyepatch, after that you can go with her default bodysuit, skin tight, or play with her clothing, style, or pose.Known Issues:<ul><li>When she's not given a clothing prompt the generation may appear blurry, please remember to add as much detail as possible.</li><li>It has weird leg amputation problems sometimes, I recommend using "bad legs, amputee" in negative prompts or even better using a bad prompt embedded of your choice. Sometimes can react positively to leg-related clothing like "jeans" or "yoga pants", and can bypass this problem.</li><li>Her eyepatch is still not that accurate but it's good enough.</li></ul>old Nu-13low and Nu-13high: Both versions trigger with Nu13, but it is also better to add ahoge, eyepatch, red eyes, single braid for better results, bodysuit depends if you want the default clothing, both white and silver hair prompts works for default.Really recommended to add ahoge because text generation tends to confuse ahoge with animal ears without prompt.

Nu-13 (Blazblue)

Training res went from 512 to 768 for better consistency on details like the eyepatch. Removed the Lambda-11 images from the training database so that when generating bodysuits doesn't render them with a black color.

Files with SHA256: 50cb0e35b5430424ec7bce796ae3150a921c3e43556846fb3dcde519572ae1b3