Fine-grained accelerators for sparse machine learning workloads | IEEE Conference Publication | IEEE Xplore