Learning both Weights and Connections for Efficient Neural Networks

Song Han, Jeff Pool, John Tran +1

Neural networks are both computationally intensive and memory intensive, making them difficult to deploy on embedded systems. Also, conventional networks fix the architecture befor…