LORA Concept Group MIX

其实，我不知道我应该混合哪些模型，你能感觉到的，我缺乏想象力。

Actually, I don't know what models I should mix because I lack imagination. You can tell that, right?

好了这是一个新的合集，原先的合集上传失败了好几次，再加上狗屎的现实生活，我的心情实在是低落。所以我想讲述一些没人喜欢听的故事，我自己的融合经历，以及分享一些发现。最后是未来我模型混合的方向。

This is a new collection. The original collection failed to upload several times, and coupled with the shitty reality of life, my mood is really low. So I want to tell some stories that no one likes to hear, my own fusion experience, and share some discoveries. Finally, it is the direction of the future fusion of my model.

我在模型混合时发现了一个无聊的现象，它很常见，你们混合时肯定也遇见过。当然我并不想讲明这个现象，因为那样做很无趣。

那个现象可以大幅度调整画面效果，当我发现这个特点后我便开始进行研究，我早期上传的模型都是基于这一点的探索。

而在混合PVC模型期间我发现了一个有趣的特点，并尝试提取概念特征进行混合。但是，在这一个新领域里，你很难验证自己的行为和猜想，也无从得知自己的行为是否正确。总之我尝试制作了两个概念提取模型。

早期的模型都有一个缺点那就是缺乏独特性。而基于画面调整的模型混合也走向了末路，因为这始终无法模拟艺术家画风。于是我不得不进行下一步的混合，通过那个现象进行特征过滤，然后提取，或许我验证了对于模型概念的猜想，但我并不确定。

至于那个现象，你只需要进行足够多的混合就会发现，甚至讲述这个现象本身就是很无聊的。

未来我的模型混合的方向：

组建LORA概念集合然后进行混合

收集一些基础画风LORA进行混合

验证一些新的想法

I discovered a boring phenomenon while mixing models. It is very common, and you must have encountered it when mixing. Of course, I don't want to explain this phenomenon because it's not interesting.

That phenomenon can greatly adjust the visual effect. When I discovered this feature, I started to conduct research. The models I uploaded earlier were all based on this exploration.

During the process of mixing PVC models, I discovered an interesting characteristic and attempted to extract conceptual features for mixing. However, in this new field, it is difficult to verify one's actions and hypotheses, and it is impossible to know whether one's actions are correct. In any case, I attempted to create two conceptual extraction models.

Early models all had a common flaw, which was a lack of uniqueness. And the blending of models based on image adjustment also reached a dead end, because it could never simulate the artist's painting style. So I had to proceed to the next step of blending, filtering through that phenomenon, and then extracting. Perhaps I have validated my conjecture about the concept of the model, but I am not certain.

As for that phenomenon, you will only need to mix enough to discover it, even telling about the phenomenon itself is very boring.

The direction of my future model MIX：

Build a LoRa concept collection and then mix it.

Collect some basic art styles of Lora and mix them together.

Validate some new ideas.