Suggested weight: 0.8
Recommended model: Counterfeit v30
This LoRA training uses the full tag mode, so it needs more character description tag when calling
The next version will test and adjust the fit of LoRA
建议权重:0.8
推荐模型:Counterfeit v30
本LoRA训练时采用全tag的方式,所以调用时需要更多的人物描述tag
下一个版本将会对LoRA的拟合度进行测试与调整