Hiddenite: 4K-PE Hidden Network Inference 4D-Tensor Engine Exploiting On-Chip Model Construction Achieving 34.8-to-16.0TOPS/W for CIFAR-100 and ImageNet

📋 论文概要

该论文提出名为Hiddenite的4K处理单元(PE)的隐藏网络推理4D张量引擎，通过利用片上模型构建实现稀疏子网络的高效推理。解决了传统密集神经网络计算量大、资源消耗高的问题，同时保持了与原始密集模型相当的精度。

💡 主要创新点

• 提出基于Hidden Network (HNN)训练方法的片上稀疏模型构建技术，实时提取高效子网络。
• 设计4K-PE的4D张量引擎架构，支持高并行度的稀疏矩阵运算，实现34.8-16.0 TOPS的峰值性能。

🏷 关键词

📄 原文摘要

Ángel López García-Arias, Junnosuke Suzuki, Thiem Van Chu, Kazushi Kawamura, Masato Motomura Tokyo Institute of Technology, Yokohama, Japan *Equally Credited Authors (ECAs) Since the advent of the Lottery Ticket Hypothesis![1], which advocates the existence of embedded sparse models that achieve accuracies equivalent to the original dense model, new algorithms to find such subnetworks have been attracting attention. In particular, Hidden Network (HNN)! [2] proposed a training method that finds such accurate subnetworks (Fig.!15.4.1). HNN extracts the sparse subnetwork by taking a logical AND of an initial model’s random weights and a binary mask that defines the selected connections – a supermask. The importance of each connection, quantified as a score, is evaluated in the training phase; a supermask is learned by picking the connections with the top-k% highest scores. Although similar to pruning, supermask training is clearly

👥 作者与机构

Kazutoshi Hirose*, Jaehoon Yu*, Kota Ando, Yasuyuki Okoshi,