STEP1: Modifying Makefile.config in Caffe-master
Tip: You can go to the file directly "Ctrl+f", type the appropriate capital letter to find the appropriate location.
STEP2: Compile interface. If the previous compilation Caffe environment, has been runtest, you can directly compile the interface, open the terminal in the Caffe-master directory, enter: Make Catcaffe.
Note: Be sure to remember before this step to modify the address in the makefile, as for if the CPU version, remember to go "only CPU" "#", I believe in the configuration of the environment has been noted.
STEP3: Test interface. Input: Make Mattest
If the error here can refer to http://www.cnblogs.com/laiqun/p/6031925.html
If it works, it will usually appear as follows:
STEP4: Can download a trained network to test the interface.
A. You can download the appropriate training network to Http://dl.caffe.berkeleyvision.org/bvlc_reference_caffenet.caffemodel and put it in the folder/caffe-master/ Models/bvlc_reference_caffenet, this is a convenient model for running demo use.
B. In terminal input "matlab" to bring up matlab2015b, switch to directory/caffe-master/matlab/demo/(this is important)
C. Enter the command run (' CLASSIFICATION_DEMO.M ') or double-tap open CLASSIFICATION_DEMO.M directly click "Run" on the console above, no input parameters required.
D. The output is a 1000x1 matrix because the Imagenet dataset has 1000 categories.
To the end of this test. Good luck to you all!
ubuntu14.04&matlab2015b Test Caffe Matlab interface