I increased the number of training images to 90, which produced an overtrained version of LoRA. So I mixed it with the original version 0.4 to produce this version
Better interaction with other LoRA's.
Better understanding of concepts.
More facial features for characters
Can be used with more realistic models (at higher value)
I increased the number of training images to 90, which produced an overtrained version of LoRA. So I mixed it with the original version 0.4 to produce this version
Better interaction with other LoRA's.
Better understanding of concepts.
More facial features for characters
Can be used with more realistic models (at higher value)