Layer Folding: Neural Network Depth Reduction using Activation Linearization

Re-assign community