paper-with-me

홈 › Papers

Algorithm-Hardware Co-Design of Distribution-Aware Logarithmic-Posit Encodings for Efficient DNN Inference

2024-03-08 · Akshat Ramachandran, Zishen Wan, Geonhwa Jeong, John Gustafson, Tushar Krishna

Traditional Deep Neural Network (DNN) quantization methods using integer, fixed-point, or floating-point data types struggle to capture diverse DNN parameter distributions at low precision, and often require large silicon overhead and intensive quantization-aware training. In this study, we introduce Logarithmic Posits (LP), an adaptive, hardware-friendly data type inspired by posits that dynamically adapts to DNN weight/activation distributions by parameterizing LP bit fields. We also develop a novel genetic-algorithm based framework, LP Quantization (LPQ), to find optimal layer-wise LP parameters while reducing representational divergence between quantized and full-precision models through a novel global-local contrastive objective. Additionally, we design a unified mixed-precision LP accelerator (LPA) architecture comprising of processing elements (PEs) incorporating LP in the computational datapath. Our algorithm-hardware co-design demonstrates on average <1% drop in top-1 accuracy across various CNN and ViT models. It also achieves ~ 2x improvements in performance per unit area and 2.2x gains in energy efficiency compared to state-of-the-art quantization accelerators using different data types.

📄 PDF Abstract BibTeX arXiv:2403.05465

Code (1)

georgia-tech-synergy-lab/LogarithmicPosit 공식 구현

Tasks

Quantization

Similar Papers 제목 키워드 기반

Power-of-Two Quantization for Low Bitwidth and Hardware Compliant Neural Networks

2022-03-09 · Dominika Przewlocka-Rus, Syed Shakib Sarwar, H. Ekin Sumbul, Yuecheng Li 외

Deploying Deep Neural Networks in low-power embedded devices for real time-constrained applications requires optimization of memory and computational complexity of the networks, usually by quantizing the weights. Most of…

Quantization

Hardware Complexity Aware Design Strategy for a Fused Logarithmic and Anti-Logarithmic Converter

2020-11-12 · Botao Xiong, Yuanfeng Sui

The logarithmic and anti-logarithmic converters are realized with the piecewise linear approximation method, which is implemented by the shift-and-add architecture. This brief utilizes the similarities of Log and Antilog…

Automated Log-Scale Quantization for Low-Cost Deep Neural Networks

2021-06-19 · CVPR 2021 1 · Sangyun Oh, Hyeonuk Sim, Sugil Lee, Jongeun Lee

Quantization plays an important role in deep neural network (DNN) hardware. In particular, logarithmic quantization has multiple advantages for DNN hardware implementations, and its weakness in terms of lower perform…

Image EnhancementQuantizationSemantic Segmentation

A Time-to-first-spike Coding and Conversion Aware Training for Energy-Efficient Deep Spiking Neural Network Processor Design

2022-08-09 · Dongwoo Lew, Kyungchul Lee, Jongsun Park

In this paper, we present an energy-efficient SNN architecture, which can seamlessly run deep spiking neural networks (SNNs) with improved accuracy. First, we propose a conversion aware training (CAT) to reduce ANN-to-SN…

Bandit algorithms: Letting go of logarithmic regret for statistical robustness

2020-06-22 · Kumar Ashutosh, Jayakrishnan Nair, Anmol Kagrecha, Krishna Jagannathan

We study regret minimization in a stochastic multi-armed bandit setting and establish a fundamental trade-off between the regret suffered under an algorithm, and its statistical robustness. Considering broad classes of u…