Optimal channel selection with discrete QCQP

Yeonwoo Jeong, Deokjae Lee, Gaon An +2

Reducing the high computational cost of large convolutional neural networks is crucial when deploying the networks to resource-constrained environments. We first show the greedy ap…