Style LoRA trained on on NoobAI-XL vpred 1.0 with images from the primarily NSFW artist D-ART:<a target="_blank" rel="ugc" href="https://twitter.com/18dart5">https://twitter.com/18dart5</a> Trigger word is jz235 It works on mentioned model and its mixes: <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294/noobai-xl-nai-xl">https://civarchive.com/models/833294/noobai-xl-nai-xl</a> Negatives and quality tags are optional. CFG around 3-4 is enough. v2.0 is new and more faithful to D-Arts's style and in general seems to have better anatomy. All previous watermarks that appeared in the corners of images are no longer there (i removed them from the training images. I recommend v2.0 as the model of choice. v1.0 is smoother and more clean in looks but is underfit and has a hard time with applying it's style on characters that are overfit on NoobAI (like Asuka (soryuu asuka langley)) and has worse anatomy.

D-ART Artist Style LoRA (Noob Vpred)

<ul><li>MUCH larger dataset (85 images) - first autotagged and then manually checked, removing unnecessary tags/adding missing tags to images as needed.</li><li>ALL watermarks removed from training images (yes, ALL. Thanks IOpaint!)</li><li>pre-downscaled images before training using custom image downscaling script<ul><li>sd-scripts does this already, but I wanted to achieve a higher downscale quality, using a script that downscales the images using a combination of Lanczos (40 %) and LSID (60%) algorithm by blending images from both algorithms together at mentioned percentages.<ul><li>LSID from paper: <a target="_blank" rel="ugc" href="https://arxiv.org/abs/2510.24334">https://arxiv.org/abs/2510.24334</a></li></ul></li></ul></li></ul><ul><li>using edm instead of minsnr for timestep correction/training</li><li>final lora model is a 50/50 blend of epoch 25 and epoch 35, using the lora-merger in sd-scripts.</li><li>resized lora using this script, with a threshold of 2.85, shaving off 30-35 MBs and reducing overfitting:<ul><li><a target="_blank" rel="ugc" href="https://github.com/elias-gaeros/resize_lora">https://github.com/elias-gaeros/resize_lora</a></li></ul></li></ul>

Files with SHA256: b48b528c6e2b54bbdceb7e7c27b6f77e9f580451dd6f964a64e949f5fce0e8e7