FaCT-LSTM: Fast and Compact Ternary Architecture for LSTM Recurrent Neural Networks
- 31 March 2021
- journal article
- editorial
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Design & Test
- Vol. 39 (3) , 45-53
Abstract
This article proposes a Fast and Compact Ternary LSTM (FaCTLSTM), which bridges the accuracy gap between the full-precision and quantized neural networks.Keywords
This publication has 8 references indexed in Scilit:
- MuBiNN: Multi-Level Binarized Recurrent Neural Network for EEG Signal Classification2020
- Universal Deep Neural Network CompressionIEEE Journal of Selected Topics in Signal Processing, 2020
- Multi-level Binarized LSTM in EEG Classification for Wearable Devices2020
- TOT-Net: An Endeavor Toward Optimizing Ternary Neural Networks2019
- Accelerating Inference In Long Short-Term Memory Neural Networks2018
- XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks2016
- LSTM: A Search Space OdysseyIEEE Transactions on Neural Networks and Learning Systems, 2016
- Low-Power ECG-Based Processor for Predicting Ventricular ArrhythmiaIEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2015