Flattening Sharpness for Dynamic Gradient Projection Memory Benefits Continual Learning Danruo Deng

Open in new window