id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_ie6u4zqq7vhp3k4oezpc3koj2q	Haoqi Yang	Street View House Number Identification Based on Deep Learning	2019	6	.pdf	application/pdf	2797	212	58	Street View House Number Identification Based on Deep the network optimizes the input layer to accept three-channel number(SVHN) dataset, effectively improving the performance number, use the extracted features to train the SVM The samples in the data set are singleInternational Journal of Advanced Network, Monitoring and Controls Volume 04, No.03, 2019 optimizer has almost no improvement in the test set of the test set except the SGD optimizer is not accuracy of the test set of the other three networks even the highest test set accuracy, only the 7th epoch highest test set accuracy of the network using the other Comparing the training set accuracy rate of the test set in Figure 5, the curve with the weight view number, the better public data set is the SVHN extended set to the model training. Therefore, the different training sets in Figure 13 SVHN dataset to improve the classic LeNet-5 network 	./cache/work_ie6u4zqq7vhp3k4oezpc3koj2q.pdf	./txt/work_ie6u4zqq7vhp3k4oezpc3koj2q.txt