TopoNets: High Performing Vision and Language Models with Brain-Like Topography