qian feng

Incremental Learning of Vision-Language Models via Task Subspace Projection and Dynamic LoRA

LW2G: Learning Whether to Grow for Prompt-based Continual Learning