I noticed that the model is quite bit noisy and bit blurry. Here's some changes in the model:
All dataset increased its resolution by 4x from v1.0. This change hopefully makes model has much higher quality and fidelity due to higher resolution
LoRA's size increased
Optimization in pruning some dataset by cropping unnecessary part of images
Fixed issue when the subject didn't have correct hair color (Instead of green, gets black haired)
I noticed that the model is quite bit noisy and bit blurry. Here's some changes in the model:
All dataset increased its resolution by 4x from v1.0. This change hopefully makes model has much higher quality and fidelity due to higher resolution
LoRA's size increased
Optimization in pruning some dataset by cropping unnecessary part of images
Fixed issue when the subject didn't have correct hair color (Instead of green, gets black haired)