spatial dropout
Description: Spatial dropout is a neural-network regularization method that drops entire feature channels or spatially correlated activations during training.
Related topic: Advanced Neural Network Applications (19%)
OpenAlex ID: keywords/spatial-dropout
