My own trained Lora of character Mona from Genshin Impact, I think it's good to have options. This is trained on the og anime model NovelAI (animefull-final-pruned) in order for the Lora to be as flexible as posible. I've trained it on 40 high quality images hand picked by me and trained at 768x768. It was trained on Clip skip 2. So don't forget to set it in your settings before using this Lora for optimal results. As trigger words, besides the main trigger word (mona_tpa), use witch hat, pantyhose and/or star_(symbol) pantyhose as well for her signature default attire. As mentioned above since it was trained on NovelAI model it means it works great on checkpoints based on this model (which are plenty). As you will see in the examples provided, it's very easy to change her default outfit. When changing her default outfit sometimes stuff like her hair color or eyes tend to be slighty different than the original. Just use "purple hair" and "aqua eyes" to guide the AI back to her design. And ofcourse, NSFW because we're all a bunch of degen.. I mean man and women of culture. I sprinkled NSFW images in the training as well (25% of the images), so you can go to town with NSFW as well. Like always with Loras, 1.0 can be overpowering so use weights like 0.7-0.9 for best results. <h3>Checkpoints used in the examples: </h3><ul><li>AbyssOrange 2 Hardcore </li><li>AbyssOrange 2 NSFW </li><li>AnyHentai </li><li>Anything 3.0 </li><li>Grapefruit </li><li>KotosMix </li></ul>Don't limit yourself only to these, any model based on NAI should produce nice results. Feel free to share your creations

Mona Genshin Impact by Kangaxx

Mona Genshin - Kangaxx.safetensors

v1.0

原作者说明：我自己训练的 lora 角色莫娜来自原神，我认为有选择是件好事。这是在 og 动漫模型上训练的 NovelAI （animefull-final-pruned） 为了使 Lora 尽可能灵活.我已经在我精心挑选的 40 张高质量图像上训练了它，并以 768x768 进行了训练。它是在Clip skip 2 上训练的。所以不要忘记在使用此 Lora 以获得最佳效果之前在您的设置中进行设置.作为触发词，除了主要的触发词（mona_tpa）外，还使用witch hat, pantyhose或star_(symbol) pantyhose作为她的标志性默认服装。如上所述，由于它是在 NovelAI 模型上训练的，这意味着它在基于该模型的检查点（很多）上效果很好。正如您将在提供的示例中看到的，更改她的默认服装非常容易。在改变她的默认服装时，有时她的头发颜色或眼睛等东西往往与原来略有不同。只需使用“purple hair 紫色头发”和“aqua eyes 水绿色眼睛”即可引导AI回到她的设计中。当然，还有NSFW，因为我们都是一群堕落者。我的意思是文化的男人和女人。我在训练中也撒上了NSFW图像（占图像的25%），所以你也可以和NSFW一起去镇上。像 Loras 一样，1.0 可能会压倒一切，因此请使用 0.7-0.9 等权重以获得最佳效果。示例中使用的检查点：AbyssOrange 2 ********AbyssOrange 2 NSFWAnyHentaiAnything 3.0GrapefruitKotosMix不要将自己局限于这些，任何基于 NAI 的模型都应该产生不错的结果。随时分享您的创作

Mona Genshin 莫娜（原神）

mona_tpa, 1girl, witch hat, stairs, shrine, lens flare, star_(symbol) pantyhose, day, aqua eyes, small breasts, cloudy sky, highly detailed, absurdres, masterpiece, highly detailed, absurdres, masterpiece, 

Mod from <a target="_blank" rel="noopener noreferrer nofollow" href="https://civitai.com/models/43201/mona-genshin-impact-by-kangaxx">Mona Genshin Impact by Kangaxx - v1.0 | Stable Diffusion LoRA | Civitai</a>Trigger WordsMONA_TPAWITCH HATPANTYHOSESTAR_(SYMBOL) PANTYHOSEMy own trained Lora of character Mona from Genshin Impact, I think it's good to have options.This is trained on the og anime model NovelAI (animefull-final-pruned) in order for the Lora to be as flexible as posible.I've trained it on 40 high quality images hand picked by me and trained at 768x768. It was trained on Clip skip 2. So don't forget to set it in your settings before using this Lora for optimal results.As trigger words, besides the main trigger word (mona_tpa), use witch hat, pantyhose and/or star_(symbol) pantyhose as well for her signature default attire.As mentioned above since it was trained on NovelAI model it means it works great on checkpoints based on this model (which are plenty).As you will see in the examples provided, it's very easy to change her default outfit. When changing her default outfit sometimes stuff like her hair color or eyes tend to be slighty different than the original. Just use "purple hair" and "aqua eyes" to guide the AI back to her design.And ofcourse, NSFW because we're all a bunch of degen.. I mean man and women of culture. I sprinkled NSFW images in the training as well (25% of the images), so you can go to town with NSFW as well.Like always with Loras, 1.0 can be overpowering so use weights like 0.7-0.9 for best results.Checkpoints used in the examples:<ul><li>AbyssOrange 2 Hardcore</li><li>AbyssOrange 2 NSFW</li><li>AnyHentai</li><li>Anything 3.0</li><li>Grapefruit</li><li>KotosMix</li></ul>Don't limit yourself only to these, any model based on NAI should produce nice results.Feel free to share your creations

Files with SHA256: ab6a06398f44e74fef6b70f1fa8f8e70c6891846e28cbce80171b717b882f8f8