Calibration-free compression brings Evo 2 to its full million-token context on a single GPU
Michail Patsakis, Alexandros Tzanakakis, Ilias Georgakopoulos‐Soares
The paper introduces TurboQuant-Bio, a toolkit that compresses Evo 2's weights and attention cache to four bits without calibration data, enabling genome-scale analysis on single GPUs.