Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning | Read Paper on Bytez