上一条： Esnreram: An energy-efficient sparse neural network based on resistive random-access memory
下一条： GPNPU: enabling efficient hardware-based direct convolution with multi-precision support in GPU tensor cores
