摘要
针对传统Vgg网络对煤矸石分类任务训练速度慢的问题,本文使用两种深度学习网络对煤矸石样本图像进行分类测试,并与Vgg网络进行性能对比,发现相比传统的Vgg深度学习网络,轻量化深度学习网络的模型训练速度明显加快,在保证分类精度基本不下降的前提下模型训练时间可减少75%。
In view of the slow training speed of traditional Vgg network for coal gangue classification task,this paper uses two deep learning networks to classify and test coal gangue sample images,and compares their performance with Vgg network.It is found that compared with traditional Vgg deep learning network,the model training speed of lightweight deep learning network is obviously accelerated,and the model training time can be reduced by 75%while the classification accuracy is basically not reduced.
作者
王国新
陈思羽
张冬妮
Wang Guoxin;Chen Siyu;Zhang Dongni(School of Electrical and Control Engineering,Heilongjiang University of Science and Technology,Harbin Heilongjiang,150022)
出处
《电子测试》
2021年第18期56-57,68,共3页
Electronic Test
基金
2020年度黑龙江省省属高等学校基本科研业务费科研项目(702/0000070210)。
关键词
煤矸石分类
深度学习
轻量化
训练速度
Gangue classification
deep learning
lightweight
training speed