4 ms·Network to Network Compression via Policy Gradient Reinforcement Learning1 points by machinelearning 9y ago