“也许,兵马俑可以和机甲组合一下,来点大秦的震撼”
由多张兵马俑图片包括概念图广告等训练。无提示词,强烈风格震撼。请注意,本lora可能与一些纯2d模型相性较差。
示例图片前两张是简单tag只加本lora的效果,
后面几张是多个lora同时作用,部分图使用了控制网的效果,仅供参考。
ღ( ´・ᴗ・` )比心
Steps: 10, Sampler: DPM++ 2S a Karras, CFG scale: 7, Size: 512x640,
Hires upscale: 2, Hires steps: 10, Hires upscaler: R-ESRGAN 4x+
以下是机翻
====================================
"Perhaps, the terracotta warriors and horses can be combined with mecha, to the shock of the Qin Dynasty."
By a number of terra-cotta pictures including concept images advertising and other training. No hint words, strong style shock. Note that this lora may be poorly compatible with some pure 2d models.
The first two sample pictures are simple tag with the effect of adding only this lora.
The next few pictures show multiple lora working at the same time, and some of the diagrams use the effect of the control network for reference only.
ღ( ´・ᴗ・` )比心
Steps: 10, Sampler: DPM++ 2S a Karras, CFG scale: 7, Size: 512x640,
Hires upscale: 2, Hires steps: 10, Hires upscaler: R-ESRGAN 4x+
Description
FAQ
Comments (26)
好活
大写的服!
看起来非常忠诚🤣
雀食震撼
可以的,少年!
前卫就是复古!
新时代动力装甲
Great job ! Just 1 full prompt and the recommended models ?
I didn't use the exact cue words when training, in fact this style is global coverage,
Model: abyssorangemix3AOM3_aom3a3,
ღ( ´・ᴗ・` )
@timevisitor So how to call/include this LORA from the prompt ?
@ritcher1 You just add it, and if it's not obvious you can add a "statue"
ok, no special keywords ... almost all the loras are called in the prompt with something like this:
<lora:meiOverwatchLORA_v1:0.8>人才
用这个LORA必须给秦始皇V50
厉害呀,这个太喜欢了。
真的是太屌了
插个眼,好活!
秦始皇都得给你V50
牛B,震惊了。
秦始皇の赛博兵马俑
逆天!
能做个三星堆的嘛。我想看看效果。






