
啊,用光辉当底模重置了这个,原版的训练集被我删掉以一部分,原因是在马斯克宣布发布在X上的图片自动同意运行ai训练以后,许多画师把自己的作品从x上删掉了,所有我训练的时候也删掉了,用法和光辉一样,更擅长画风景,
Ah, I reset this using Illustrious as the base model. Part of the original training set was deleted because, after Elon Musk announced that images posted on X would automatically consent to being used for AI training, many artists removed their works from the platform. So, I also deleted them during training. It functions the same way as Illustrious and is better at drawing landscapes.
以下是第一次更新的内容
Below is the content of the first update
我暂时还没有在C站上看到类似的画风,所以就上传了,虽然这是个融合模型,但是融合素材都是我自己制作的,
I haven't seen a similar painting style on Civital yet, so I uploaded it.Although this is a fusion model. But the fusion materials are all made by myself.
只要你写了的tag!就会有回响! 但是只有你写了的tag!才会有回响!!! 所以不要问为什么别的模型能出这个模型不能出!你先看看你写没写!!!或者写了什么奇怪的东西!!!
描述:
一个泛化的大模型(不够二次元真是不好意思啊!),相比于之前的千界婆娑,更加的干净,画面更加清爽,污染更小(整天就为这些事情劳心劳力了,新版的千界婆娑是可以画萝莉的!)。
可以兼容各种lora,之后例图上使用过的lora能上传的我也会尽量上传。
(目前1.X很多事情XL还是做不到的个说)。
最佳分辨率其实是640x960加2倍的高清()这是我平时跑图的设置。
喜欢的话请给我支持,返图也好,收藏也好,最近有些emo,需要大家的鼓励(我在说什么啊!)
训练词语:
名称: shyakujxmix_V2.safetensors
大小 (KB): 4165134
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success