Ethereal Eastern -Flux（Digital painting+ink freehand brushwork）

MIX

A fusion version of V1 and V2.

This is a trial version, V2 will be more stable, but this version seems to provide more diversity.Effective weight:0.8-1.5,recommendation 1.35-1.5.Trigger word: Ink freehand brushwork style

————————

V1和V2的融合版本。这是个试验版本，V2会更稳定，但这个版本似乎能提供更多的多样性。

有效重量:0.8-1.5，建议1.35-1.5。触发词：水墨写意风格

V2

I expanded my photo collection and trained women's faces and overall painting style. Compared with the previous version, this version is more prominent in the brushwork of ink and wash and the freehand brushwork style of Chinese painting.

The weight is relatively stable between 0.5 - 0.8. I usually start using it around 0.6, and Then adjust it step by step according to the effect in the unit of 0.05. Generally speaking, between 0.6 and 0.8, the picture will be close to what I am trying to achieve, that is, the combination of more flexible oriental ink painting style and digital painting.

You can also use this lora as a pepper seasoning. For example, if you want Spider-Man, Superman or Batman to present this modern ink painting style, you can create pictures with a weight of 0.45 or even lower.

I have tried 0.3, in fact, it has a unique charm, it is closer to realism, but at the same time there will be some performance in painting style. Welcome to try ~ You can use [Ink freehand brushwork style] as a trigger word, but it may be better to use natural language.

______________

我拓展了图片集，对女性的面部和整体画风进行了针对化的训练。相比上一个版本，这一版本在水墨笔触和中国画的写意画风上更加突出。

权重在0.5-0.8之间相对都是稳定的，我一般会在0.6左右开始使用它，然后根据效果按照0.05为单位逐步去调整。通常来说，0.6-0.8之间，画面会靠近我试图达到的效果，也就是较为灵动的东方水墨风格和数字绘画的结合。

你也可以将这个lora作为胡椒粉一样的调料来使用，比如你想让蜘蛛侠、超人或者蝙蝠侠呈现这种现代水墨画风，你可以在0.45甚至更低的权重下去生图。

尝试过0.3，实际上它别有一番韵味，它更靠近写实，但同时也会有一些画风上的表现。欢迎去尝试~

你可以使用【Ink freehand brushwork style】作为触发词，但使用自然语言的效果也许更好。

_________________________________

V1

Description:

Just like the name of this lora-Ethereal Eastern-this is a lora that adds a smart style to the Flux model, and it has a strong oriental color.You can understand it as a fusion experiment of digital painting and oriental ink painting style.

Since the release of Flux, I have generated a lot of pictures with it, and its power is amazing. But the fly in the ointment is that I found that the pictures generated by Flux are relatively "realistic". What we refer to here is not the content it generates, but the aesthetic tendency of the picture it generates, which includes techniques such as lines and colors. Most of the pictures we get from Flux are "concrete", "thick" and "accurate", and sometimes, this lacks some light aesthetic feeling. Just like some oil paintings, they are always infinitely close to the truth. So I trained this Lora.

At first, I just wanted to train an ink-and-wash lora and generate some characters similar to splash-ink style, but later I changed my goal. I collected some pictures according to my own aesthetics, which are not limited to ink-and-wash style. Generally speaking,I hope that these pictures can be integrated into oriental aesthetics while having modern digital painting style. Most of the pictures have a sense of freehand brushwork in ink and wash, Some lines are relatively light and agile, and tend to be meticulous in Chinese painting, while some pictures have watercolor style.

But most of the pictures are very modern in aesthetics, mostly illustrations and cartoons. It is worth mentioning that some of the pictures in the photo collection are from midjourney. Because this kind of style pictures of midjourney are excellent, it is almost unavoidable. Correspondingly, the final generated images are similar to them.

Weight:

The weight between 1.2 - 1.5 is the best, usually 1.3 is enough.

But if you want to draw some conceptual characters, such as Superman, Spider-Man and Batman, because Flux itself has a deep brand on these characters, you need to adjust the weight to around 1.4 to offset the partial realistic style of Flux.
______________________________________________________________________

说明：

就如同这个lora的名字——灵动东方——这是个给Flux模型添加灵动画风的lora，并且它带有浓郁的东方色彩。你可以理解为它是数字绘画+东方水墨风格的一次融合试验。

自从Flux发布以来，我用它生成了大量的图片，它的强大令人惊叹。但美中不足的是，我发现Flux生成的图片相对还是偏向于“写实”。

这里指的并不是它生成的内容，而是它生成画面的美学倾向，包含了线条、色彩等技法，我们从Flux得到的图片绝大多数显得“具体”“厚实”“精准”，某些时候，这缺乏了一些轻灵的美感。就像有些油画，总是无限地向真实靠近。所以我训练了这个Lora。

一开始，我只是想训练一个水墨风格的lora，生成一些类似于泼墨风格的人物，但后来我改变了目标，我按照自己的审美搜集了一些图片，它们并不局限于水墨风格，总体来说，我希望这些图片在具备现代的数字绘画风格的同时，能够融入东方美学。其中一大部分图片带有水墨写意的笔触感，部分的线条相对轻巧灵动，偏向中国画的工笔，部分图片则带有水彩风格。但大部分的图片审美非常现代，多是一些插画和漫画。值得一提的是，图片集中有一部分图片是来自于midjourney，因为midjourney的这类风格图片非常出色，所以这几乎是无法规避的，相应的，最后生成的图像，也与之有一定的相似性。

权重：

权重在1.2-1.5之间表现是最好的，通常使用1.3即可。

但如果你要画一些概念性的人物，比如超人、蜘蛛侠、蝙蝠侠，因为Flux本身对这些人物有较深的烙印，你需要把权重调到1.4左右，来抵消Flux自带的偏写实风格。