Thanks for sharing your code.

I finished training using caffe but after that i don't know how to test my data.
In SRCNN, they has matlab code for extracting parameter of caffemodel. Am i made my own code like that? How can you extract parameter from your caffemodel?

dear zehao:
About the dictionary Dx , I wonder is it pretrained or training with the network?
In addition, the paper mentioned W an S must satisfy the relationship: W=CD^(T) and S=I-D^(T)D. In issue #3, you mentioned that W and S are learned by back-propagation algorithms, I wonder if I should set corresponding weight relationship of W, S, D in training ? If the corresponding weight limit is not added in training，W, S can satisfy the relationship: W=CD^(T) and S=I-D^(T)D ?
Thank you very much

Hello! I am very interested about this project. But I don't fully understand the training detail of this algorithm. Could you share the caffe training code of SCN recently? Thank you very much.

Dear zehao
Firstly, thanks very much for the great test code of the paper. As you implemented the training code by caffe, would you please share it on github? If it contains some bugs, would you please send me a piece as I'm very interested about the paper. It will give me great help for understanding the paper and doing some future work. Thank you very much!@huangzehao

Dear zehao
I'm so glad to read this paper. I think this paper has made a great combination of domain expertise（sparse representation）and deep network. But I wonder how to train the network as there are three linear layers. What I understand about the neural network is that each layer outputs the feature map, so how can the coefficients of sparse can be output. I would appreciate it if you can provide me with the source code. Thank you again. @huangzehao

如何针对我自己的数据，进行.mat的训练，希望作者开源train.m ,test.m, SCN_solver.prototxt.
希望作者可以提供 上述文件，即双字典稀疏编码 时对应的Deep 网络结构，同时渴求作者联系邮箱，小弟有诸多SR的问题想要请教，如作者看到之后及时反馈，小弟将不胜感激。鄙人联系方式 [email protected] ，跪求大神指导。