新版本运用了分层融合的插件用来优化部分情况下的肢体崩坏情况,削减了LORA自带的画面背景污染,让画面和构图更加可控。并且这个版本不需要唤醒词,解决了编写单独tag的麻烦。
This version uses a layered fusion plug-in to optimize the limb collapse in some cases, reduces the background pollution of the picture that comes with LORA, and makes the picture and composition more controllable. And this version does not require a wake word, which solves the trouble of writing a separate tag.
这次改良的模型可以不用频繁的调整CFG值,LORA的模型权重相比上个版本也更加自由了。
The improved model this time does not need to frequently adjust the CFG value, and the model weight of LORA is more free than the previous version.
以下是我的分层融合参数(Below are my layered fusion parameters):
hipoly_3dcg_v7-epoch-000012:0.3:1,1,1,1,1,1,1,1,0,0,0,1,1,1,1,1,1
heitagirl_v3:1:1,0,0,0,0,0,0,0,0.8,1,1,0.2,0,0,0,0,0