2 LORAs trained in the style of bluefield

First version (soft), softer style - Trained less (32/16) DIM/ALPHA)

Second version (hard), harder style - Trained more (128 DIM/ALPHA)