Model Merging

Foldable SuperNets: Scalable Neural Merging with Different Initialization and Tasks