A full fine-tune aimed at improving the base model’s structure and consistency without introducing a heavy style bias.
f2k4BRefinedBase_alpha1.safetensors