Flex-sweep 2.0: more flexible and faster selective sweeps detection

Flex-sweep uses a convolutional neural network to identify selective sweeps, including ancient ones thousands of generations old, from single-population data, and remains robust to background selection. Version 2.0 streamlines the entire workflow, drastically cutting memory requirements and accelerating summary-statistic calculation over fully customizable statistic combinations and genomic regions. The update also relaxes constraints on CNN input, likely expanding compatibility with diverse data formats. These improvements make large-scale analyses of positive selection more practical, potentially enabling broader application in evolutionary genomics.