LPM-PyTorch-implementation This is a Pytorch implementation of the paper Latency-Aware Differentiable Neural Architecture Search.