The original character image in version 1.0 is more reproducible and has poor generalization.
Version 1.1
The restoration of the original image and clothing details is not as good as in 1.0. The version is more generalizable. You can wear a variety of clothes, and you can take off your naked body very cleanly. It is not like the 1.0 clothes that stick to the body and are difficult to change.
lora weight recommendation: 1 should not exceed 1, just increase the trigger words and related prompt words (it can also be reduced according to the situation)
Trigger word weight: If there are many prompt words related to the original character (headgear, hat, face, clothes), it will default to 1. If there are only prompt words related to the head and face, when using other images, it will not look like Buller at all. Can increase the weight of trigger words
