when defining new Deep learning layers, matlab do not aware my function in matlab/toolbox folder on Linux (In Macbook, works! )

Dear Matlab developer team,

Thanks for your works,

i always pleased using matlab function,

Nowadays, we are developing a new deep learning layer.

First, when you try to modify the Convolution layer and the pooling layer, you are copying existing layers and modifying only the content you want.

However, when working on a MacBook, it works well,

When I try to work on a Linux server computer, I find that I can not recognize the function and class and ask questions.

If you are developing on a Mac, can you use it to transfer to Linux?

%% examples

max_geoPooling2dLayer.m

scp -rp -P 3087 /Applications/MATLAB_R2019a.app/toolbox/nnet/cnn/*max_geo* serverID@serverIP:/usr/local/MATLAB/R2019a/toolbox/nnet/cnn

transGeometryLayer.m

scp -rp -P 3087 /Applications/MATLAB_R2019a.app/toolbox/nnet/cnn/*transG* serverID@serverIP:/usr/local/MATLAB/R2019a/toolbox/nnet/cnn

Max_geoPooling2DLayer.m

Convolution2D_geometryLayer.m

Convolution2D_geometryLayer2.m

scp -rp -P 3087 /Applications/MATLAB_R2019a.app/toolbox/nnet/cnn/+nnet/+cnn/+layer/*geo* serverID@serverIP::/usr/local/MATLAB/R2019a/toolbox/nnet/cnn/+nnet/+cnn/+layer/

TransGeometryLayer1D.m

scp -rp -P 3087 /Applications/MATLAB_R2019a.app/toolbox/nnet/cnn/+nnet/+internal/+cnn/+layer/*TransG* serverID@serverIP:/usr/local/MATLAB/R2019a/toolbox/nnet/cnn/+nnet/+internal/+cnn/+layer/

TransGeometryHostStrategy.m

scp -rp -P 3087 /Applications/MATLAB_R2019a.app/toolbox/nnet/cnn/+nnet/+internal/+cnn/+layer/+util/*TransG* serverID@serverIP:/usr/local/MATLAB/R2019a/toolbox/nnet/cnn/+nnet/+internal/+cnn/+layer/+util/

Answer 1

[Harsha Priya Daggubati]

Hi,

You can transfer the files among various OS Platforms that MATLAB supports. The issue you are facing might be because of the file permissions. You can try checking the file permissions and changing them as needed. Another reason might be due to improper installation of toolboxes on your Linux Machine.

Hope this helps!