This is a 5bpw h6 exl2 quant of Lyra4-Gutenberg2-12B
Built-in(default) calibration dataset is used.
Base model