About this document
Energy-Efficient NN Architecture for IoT by muskanshaani387 is a document available to read on EtoBox.
This document presents a low-power neural network (NN) architecture designed for error-tolerant applications, particularly in Internet-of-Things (IoT) devices. Key innovations include a dynamic reconfigurable tensor processing unit (DR-TPU) that utilizes approximate computing for energy efficiency, cross-layer error tolerance analysis, and a customized input feature map reuse strategy to minimize memory access. The proposed architecture achieves significant reductions in power consumption and enhances energ
- Author
- muskanshaani387
- Language
- EN