注意：这个模型可能只能算作一个预览版，或者直接称作实验失败品，我在训练集中加入了大量的，不同类别不同风格不同逻辑的样本，这导致难度远远超出了我的想象。这个问题我暂时无法解决，可能未来通过逐渐拆分训练集逐一训练，并且在最后合并成一个终极的Lora，才能代表这个lora的可用版本的推出，在目前为止，不推荐你在这个版本上花费太多精力来让你得到一个好的结果，但作为玩具它可以把玩，训练集中包含了大量的画师串，但我还未整理，目前这个版本在我自己的使用中，可能用于强度0.15-0.50之间，用于触发某些姿势或者画风，挺好用的。

这个lora的审美可能更偏向于男同性恋，并且训练处于初期，可能会生成非常剧烈的丑陋画面。
使得Anima的所有版本模型对某些提示词更具有响应，并赋予一定风格。对某些@画师名字可能具有反应。
由于我训练时的提示词的写法，你可能需要非常精确和没有歧义的自然语言提示词才能让其发挥良好，并且接受风格词丶质量词，如果没有使用精确丶出现歧义的提示词，画面可能崩坏或者没有顺从你的意愿。
目前的版本我在自己的使用中出现了很多不可以提示词更正的错误绘制，我目前无能为力，这个问题在以后的版本可能会慢慢进行修正。
如果你喜欢我的模型或作品，或者你手中有非常美丽的图片/训练集，你可以添加我的好友，为我提供它们，我会加入到下一次的训练，非常感谢。


Note: This model can only be regarded as a preview version, or simply called an experimental failure product. I added a large number of samples with different categories, styles, and logical structures to the training set, which resulted in a difficulty level far exceeding my expectations. I am currently unable to solve this problem. Perhaps in the future, by gradually splitting the training set and training each part separately, and then merging them into an ultimate Lora at the end, we can represent the launch of an available version of this Lora. As of now, it is not recommended that you spend too much effort on this version to obtain a good result. However, as a toy, it can be played with. The training set contains a large number of painter combinations, but I have not yet organized them. Currently, in my own usage, this version may be used within an intensity range of 0.15-0.50, to trigger certain postures or styles, and it is quite useful.

The aesthetic of this Lora may be more inclined towards gay men, and the training is at an early stage, which may result in extremely ugly images being generated.

It enables all versions of Anima's models to respond more effectively to certain prompts and give them a certain style. It may also react to certain @artist names.

Due to the way the prompts were written during my training, you may need very precise and unambiguous natural language prompts to ensure good performance, and to accept style words and quality words. If you do not use precise and ambiguous prompts, the image may collapse or not comply with your wishes.

Currently, in my own usage, there are many incorrect drawings that cannot be corrected by prompts. I am currently powerless to do anything about it. This issue may be gradually corrected in future versions.

If you like my models or works, or if you have very beautiful pictures/training sets, you can add me as a friend and provide them to me. I will be included in the next training. Thank you very much.