Version 1 was my first lora. Version 2 aims to fix the things I did wrong.
Cleaned up backgrounds of training images. Should be less artifacts.
Improved tagging of dataset
(Hopefully) improved the training parameters.
Version 1 was my first lora. Version 2 aims to fix the things I did wrong.
Cleaned up backgrounds of training images. Should be less artifacts.
Improved tagging of dataset
(Hopefully) improved the training parameters.