Few-Shot Class-Incremental Learning (FSCIL) is a novel problem setting for incremental learning, where a unified classifier is incrementally learned for new classes with very few training samples. In ...