Towards a learning-based performance modeling for accelerating Deep Neural Networks

Open in new window