Neural Network-Hardware Co-design for Scalable RRAM-based BNN Accelerators
Published in
arXiv.org(2018)External Links:
TL
TL;DR
Simulation results show that the proposed network reconstruction and retraining recovers the inference accuracy of the original BNN, and the accuracy loss of the proposed scheme in the CIFAR-10 testcase was less than 1.1% compared to the original network.
Abstract
Recently, RRAM-based Binary Neural Network (BNN) hardware has been gaining interests as it requires 1-bit sense-amp only and eliminates the need for high-resolution ADC and DAC. However, RRAM-based BNN hardware still requires high-resolution ADC for partial sum calculation to implement large-scale neural network using multiple memory arrays. We propose a neural network-hardware co-design approach to split input to fit each split network on a RRAM array so that the reconstructed BNNs calculate 1-bit output neuron in each array. As a result, ADC can be completely eliminated from the design even for large-scale neural network. Simulation results show that the proposed network reconstruction and retraining recovers the inference accuracy of the original BNN. The accuracy loss of the proposed scheme in the CIFAR-10 testcase was less than 1.1% compared to the original network. The code for training and running proposed BNN models is available at: this https URL.
References12 items
1
Input-Splitting of Large Neural Networks for Power-Efficient Accelerator with Resistive Crossbar Memory Array
2
XNOR-RRAM: A scalable and parallel resistive synaptic architecture for binary neural networks
3
Fully parallel RRAM synaptic array for implementing binary neural network with (+1, −1) weights and (+1, 0) neurons
4
FINN: A Framework for Fast, Scalable Binarized Neural Network Inference
5