优化了LORA模型,在使用realisitc模型时更加稳定。
Optimized the LORA model, which is more stable when using in the realisitc model.
recommend: majicMIX realisticV6 & XXMix_9realistic
skip clip:1
upscaler:1x_NMKDDetoon_97500_G
LoRA Strength: 0.6~0.8
size: 512 x 512,512x768
Sampler: DPM++ SDE Karras
CFG scale: 7~10
Negative embedding: EasyNegative
anime style using V1.0 setting is fine.
Model: most models are available
VAE: recommend kl-f8-anime2
LoRA Strength: 0.6~0.9(1 is also available)
size: 512 x 512,512x768
Sampler: DPM++ 2M Karras
Sampling steps: 20
CFG: 10
Negative embedding: EasyNegative
模型:大部分都可用
VAE:建议 kl-f8-anime2
LoRA强度:0.6~0.9(也可以选择1)
图片大小: 512 x 512,512x768
采样器:DPM++ 2M Karras
采样步数:20
CFG:10
Negative embedding: EasyNegative
这个LORA有四个主提示词:aura_raen,aura_xeala,black_scales,white_scales,还有可控的提示词
(晨曦):white_dhorns,white_bhorns,white_whorns,white_fhorns
(暮晖):black_dhorns,black_bhorns,black_whorns,black_fhorns
(控制4种不同的龙角生成)。
全身可以尝试下尾巴,但是不稳定,弄了很久做不出效果
尾巴提示词:
white_atail,white_btail,white_ctail,white_dtail
black_atail,black_btail,black_ctail,black_dtail
This LORA has four main trigger words: aura_raen,aura_xeala,black_scales,white_scales and controllable trigger words:
(Raen):white_dhorns,white_bhorns,white_whorns,white_fhorns
(Xeala):black_dhorns,black_bhorns,black_whorns,black_fhorns
(which control the generation of four different types of dragon horns).
You can try the tail for the full body, but it is not stable.
trigger words:
white_atail,white_btail,white_ctail,white_dtail
black_atail,black_btail,black_ctail,black_dtail
可以用:looking for viewer,from side,from behind 控制角度
You can use these prompts to control angle:looking at viewer,from side,from behind
bhorns得调高权重,不稳定,可能是哪里出问题了
bhorns have to increase the weight, not stable, maybe something is wrong
全身像可用但不稳定,建议用半身像以及头像。
Full body are available but not stable, it is recommended to use upper body and portrait.
have fun>_<