Supplementary Material for Self-Supervised Visual Representation Learning with Semantic Grouping