The data set for this model was pared down from my full model as a test model, with about ten styles built in. Currently available styles are mafuyu--e658d40fchen_bin--b19d1dd7piyodera_mucha--78a12850usashiro_mani--8d74e6c8shiratama--1dd748ecmore...(Making it public might cause unnecessary trouble)Emmm, this is just an attempt at training on Pony. The number of styles is much lower compared to my full version model (which has 500+ styles), even so, it has been trained for about 6000 steps per style. (This is about three times the training volume of the full version trained on A3.1.) Since it is a short hash, there may be hash conflicts

Pony-multi-Artist:Sytle(chen_bin,shiratama,mafuyu...)

PONY_NYA_DIFFUSION_LOW_WEAKEN-T3.safetensors

T-PONY-4.1-A1+

PONY_A1

鉴于模型的实际用途不受模型作者控制，因模型输出的图片所产生的一切后果由图片输出者自行承担。如果想画风还原请务必使用（<a rel="noopener noreferrer nofollow" href="https://tusi.cn/models/716496511741155704">https://tusi.cn/models/716496511741155704</a>）模型此lora的使用难度较高，在使用此lora时，请不要使用其他模型的提示词标准，请按照下方的使用教程进行提示词编写。否则该lora的控制性会极大的降低。使用教程起手式Best-A,【画风词】,【其他...】------------------tag格式像这样【质量词】,【画风词】,【人物描述】,【其他】例如Best-A,[tag1|tag2],character1,1girl,solo,white_hair............画师tag可自己控制权重，进行混合弱混合Best-A,(tag1:0.4),(tag2:0.5)强混合Best-A,[tag1|tag2]质量词：Best-AMedium-BLow-CBad-D画风词不公开（自己在帖子翻翻就行了——公开了又要被冲了）-----------------------如果还不会用的话去Q****926747912索要模型taglist，它会解决你的一切使用疑问兼容列表在模型简介...基于hansv35-v3,使用了18万张图片进行训练。目的是补充/增强hans-v35模型缺失的部分角色/风格。（别问我为什么不练全量微调，因为全量炸炉了！！！！！！！！！）增加了大约270位画师。增加了约30部动漫/galgame增加了部分克苏鲁/猎奇数据集增强了约1000位角色的概念在hans-v35-v3上使用此lora后可以像使用nai3一样使用已有的画师串可以进行任意的强弱混合（兼容性极强！）如果这里面没有你喜欢的画师风格请在评论区留言,下个版本将会添加你留言的画师风格！---如果您认为您是受训练数据影响的创作者之一，并且不希望AI学习您的作品，请与我联系。我将使用相应方法从模型中移除与您相关的数据。---

喵？！-多画风/概念-HANS/PONY

 The data set for this model was pared down from my full model as a test model, with about ten styles built in. Currently available styles are
mafuyu--e658d40f

chen_bin--b19d1dd7

piyodera_mucha--78a12850

usashiro_mani--8d74e6c8

shiratama--1dd748ec

more...(Making it public might cause unnecessary trouble)

Emmm, this is just an attempt at training on Pony. The number of styles is much lower compared to my full version model (which has 500+ styles), even so, it has been trained for about 6000 steps per style. (This is about three times the training volume of the full version trained on A3.1.)
Since it is a short hash, there may be hash conflicts

Files with SHA256: 4376a4846f679f753c9a90baa18a46532b6da41b6c2209be0d0e8f95f2825039