Fast-HuBERT: An Efficient Training Framework for Self-Supervised Speech Representation Learning