从给定的很多张自然图片中截取出大小为8*8的小patches图片共10000张，现在需要用sparse autoencoder的方法训练出一个隐含层网络所学习到的特征。该网络共有3层，输入层是64个节点，隐含层是25个节点，输出层当然也是64个节点了。这里只是上传了训练的部分代码，后续继续(A total of 10000 small patches pictures with size 8*8 are extracted from a lot of natural pictures. Now we need to use sparse autoencoder to train a hidden layer network to learn the characteristics. The network has 3 layers, the input layer is 64 nodes, the hidden layer is 25 nodes, and the output layer is, of course, 64 nodes. Here's only part of the training code, followed by)