Siamese Image Modeling for Self-Supervised Vision Representation Learning