loading a large trained classifier into the work space

I have trained a linear classifier on a large amount of data using tall arrays, training process takes very long time, about 6 days. i saved the trained classifier using -7.3V file format so that i can use it later as following:

save('gabor_feature_mdl_linear.mat', 'gabor_feature_mdl_linearl', '-v7.3')

the size of the trained classifier is more than 10GB. when i tried to load it to the work space it gives the following error:

"Unable to read file 'F:\gabor_feature_mdl_linear.mat'. Variable name length must be less than or equal to MATLAB maximum name length of 63 characters."

i tried to change the classifier name using a shoter name length, but the error still appears.

i donot want to train the classifer again. any help piease.

Answer 1

[Rahul]

Hi @Amelia,

In order to diagnose this issue, consider following the given steps:

• Ensure that your 'mat' file is not corrupted. Try opening it using a different machine or MATLAB version if possible.
• Secondly, try using 'matfile' function that lets you access and change variables in MAT-file without loading file into memory.
• Consider the following MATLAB Answers as well:

https://www.mathworks.com/matlabcentral/answers/2095811-unable-to-read-some-variables-error-when-converting-to-fmu

https://www.mathworks.com/matlabcentral/answers/12350-variable-name-length-exceeds-63-characters

The following MathWorks documentations can be referred:

'matfile': https://www.mathworks.com/help/matlab/ref/matlab.io.matfile.html

Thanks.