Neural network inference engine that delivers GPU-class performance for sparsified models on CPUs - View it on GitHub