Structured Sparse Ternary Weight Coding of Deep Neural Networks for Efficient Hardware Implementations | Read Paper on Bytez