Single-Model and Any-Modality for Video Object Tracking