Now showing items 1-2 of 2
Mixed-precision architecture for flexible neural network accelerators
(Massachusetts Institute of Technology, 2019)
Model quantization provides considerable latency and energy consumption reductions while preserving accuracy. However, the optimal bitwidth reduction varies on a layer by layer basis. This thesis suggests a novel neural ...
Mixed-precision NN accelerator with neural-hardware architecture search
(Massachusetts Institute of Technology, 2020)
Neural architecture and hardware architecture co-design is an effective way to enable specialization and acceleration for deep neural networks (DNNs). The design space and its exploration methodology impact efficiency and ...