Deep Learning

Foldable SuperNets: Scalable Neural Merging with Different Initialization and Tasks