期刊
ELECTRONICS
卷 11, 期 7, 页码 -出版社
MDPI
DOI: 10.3390/electronics11071101
关键词
neural architecture search; conditional CNN; genetic algorithm; performance prediction; deep learning
资金
- University of Seoul
This paper proposes a NAS technique, called CondNAS, for conditional CNN architecture. By using machine learning models and genetic algorithm, CondNAS efficiently finds a near-optimal conditional CNN architecture. The experimental results show that the conditional CNNs from CondNAS are 2.52 times faster than the CNNs from OFA on GPU and 1.75 times faster on CPU.
As deep learning has become prevalent and adopted in various application domains, the need for efficient convolution neural network (CNN) inference on diverse target platforms has increased. To address the need, a neural architecture search (NAS) technique called once-for-all, or OFA, which aims to efficiently find the optimal CNN architecture for the given target platform using genetic algorithm (GA), has recently been proposed. Meanwhile, a conditional CNN architecture, which allows early exits with auxiliary classifiers in the middle of a network to achieve efficient inference without accuracy loss or with negligible loss, has been proposed. In this paper, we propose a NAS technique for the conditional CNN architecture, CondNAS, which efficiently finds a near-optimal conditional CNN architecture for the target platform using GA. By attaching auxiliary classifiers through adaptive pooling, OFA's SuperNet is successfully extended, such that it incorporates the various conditional CNN sub-networks. In addition, we devise machine learning-based prediction models for the accuracy and latency of an arbitrary conditional CNN, which are used in the GA of CondNAS to efficiently explore the large search space. The experimental results show that the conditional CNNs from CondNAS is 2.52 x and 1.75 x faster than the CNNs from OFA for Galaxy Note10+ GPU and CPU, respectively.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据