🎞 LoRA is trained purely on anime screencaps<hr />🗝 Activation tags are in the right panel📝 Note: Blue dress -&gt; Not enough data to make the bottom part stable enough, so the frilled skirt part is more or less random.📸 Always use ADetailer to get detailed eyes🛠 Technical notes:<ul><li>LoRA shoud be compatible with Illu models &amp; NoobAI. (Not tested.)</li><li>LoRA is compatible with various styles - Lower the weight &amp; remove "anime screencap" from the prompt, if you are getting too strong anime style</li><li>Hi. res fix used on all sample images:<ul><li>Upscaler: R-ESRGAN 4x+ Anime6B</li><li>Steps: 10-15</li><li>Denoising: 0.3</li></ul></li></ul><hr />📬Leave a like, comment, review but first of all, ENJOY ;)

Files with SHA256: 7bbbdbfa4b0540697ed1089c648f00ce955fe58dec41054c85ae0cd9edc95070