Neural Networks (nntool) out of memory problem

Hi all,

I am having a problem while using the neural network toolbox (nntool) even for a small example where the input vector is p=[0 0 1 1; 0 1 0 1] and the target vector is : t=[0 0 0 1];

the message that I got is :

??? Insufficient number of outputs from right hand side of equal sign to satisfy assignment.

Error in ==> nntool at 681
      [errmsg,errid] = me.message;

com.mathworks.jmi.MatlabException: Insufficient number of outputs from right hand side of equal sign to satisfy assignment.

I'm using R2009a and have 4GB RAM Windows vista 32-bit.

I have seen a link on the support section concerning to avoid memory errors :<http://www.mathworks.com/support/tech-notes/1100/1107.html> but when I try to check on the memory as mention in the link i got:

      Physical Memory (RAM):
          In Use:                             1546 MB (60a9d000)
          Free:                               1491 MB (5d366000)
          Total:                              3038 MB (bde03000)
      Page File (Swap space):
          In Use:                             1722 MB (6ba29000)
          Free:                               4554 MB (11cadc000)
          Total:                              6277 MB (188505000)
      Virtual Memory (Address Space):
          In Use:                              640 MB (2800a000)
          Free:                               1407 MB (57fd6000)
          Total:                              2047 MB (7ffe0000)
      Largest Contiguous Free Blocks:
           1. [at 22eb6000]                   1129 MB (4693a000)
           2. [at 697f6000]                    104 MB ( 689a000)
           3. [at 7c440000]                     50 MB ( 32b0000)
           4. [at 7226f000]                     25 MB ( 1911000)
           5. [at 742ce000]                      8 MB (  882000)
           6. [at  ed50000]                      8 MB (  800000)
           7. [at 22270000]                      8 MB (  800000)
           8. [at 77b0a000]                      7 MB (  7d6000)
           9. [at 7f7f0000]                      6 MB (  69d000)
          10. [at 73dc5000]                      3 MB (  3db000)
                                              ======= ==========
                                              1352 MB (54865000

which I think is enough(at least for such an easy example) so what should I do? please tell me as it is so crucial for me sorry for being long and thanks in advance.

Answer 1

[Walter Roberson]

We need to see your code example.

The message has nothing to do with running out of memory. The message would occur if you had more variables on the left hand side of an assignment statement than were output by the expression on the right hand side of the assignment statement. For example,

[a,b] = 5

would cause the same basic problem except that the normal error message would be "Too many output arguments."

[Mostafa Mohammed]

thanks for ur reply

I was using the GUI of the (nntool) where I have created the input and the target vectors as mentioned before then i import them as input and target then i create a network and try to train it (all just using the GUI) that's when got that error message(while the training).

[Walter Roberson]

I _suspect_ you have a corrupt installation (e.g., toolbox version does not match MATLAB version), or else you happen to have a file on the path that happens to have the same name as an internal MATLAB routine.

At the command line, give the command

dbstop if error

and run again. The program will then stop in the debugger when the error is encountered. What you showed has the appearance to me that it might be from an error through from within a try/catch block; if so then you may have to trace back in the code to the beginning of the "try" portion, put a breakpoint in there, and single-step until the error occurs, and then examine in detail the operation that triggered the error.

[Mostafa Mohammed]

@ Walter

Thanks so much there were some suspicious files whom i download before and dump them to the workspace after I deleted them this small example worked. However, when I try the real required example it give me the same message !!!!!!!!!!!

the input vector dimension was 25*54000 and the target was 10*540000.

thanks for both of you any way

Answer 2

[mahdi bazarghan]

I also have the same problem, but with larger data. The code and the error massage is as follows:

>> p=load('train.dat')'; >> Tc=eye(11647); >> spread=(0.5); >> net=newpnn(p,Tc,spread); ??? Error using ==> vertcat Out of memory. Type HELP MEMORY for your options.

Error in ==> network.subsasgn>resizem at 2202 m = [m; zeros(r-R,C)];

Error in ==> network.subsasgn>setLayerSize at 1217 net.LW{i,j} = resizem(net.LW{i,j},lwSize);

Error in ==> network.subsasgn at 179 [net,err] = setLayerSize(net,i,newSize);

Error in ==> newpnn at 97 net.layers{2}.size = S;

I appreciate any help.

Cheers