This is a 6bpw h8 exl2 quant of MN-12B-Lyra-v4
Built-in(default) calibration dataset is used.
Base model