Iterative training version for full body and reducing the weight of lora.
Now, you can draw Aim's face with only 0.3 weight and leave more weight to other Lora.
When drawing a full body, Aim's face shape is maintained as much as possible, but the side length of the drawn picture cannot be less than 1024 pixels, otherwise there will not be enough pixels to draw the face (which will cause face deformation).
I suggest to keep the v2.0 version. The upper body drawn in this version is still the best.
Iterative training version for full body and reducing the weight of lora.
Now, you can draw Aim's face with only 0.3 weight and leave more weight to other Lora.
When drawing a full body, Aim's face shape is maintained as much as possible, but the side length of the drawn picture cannot be less than 1024 pixels, otherwise there will not be enough pixels to draw the face (which will cause face deformation).
I suggest to keep the v2.0 version. The upper body drawn in this version is still the best.