Within the Deep learning Toolbox it’s possible to enable the training plots within the options. It uses the following code:
Options = trainingOptions(‘Plots’, ‘training-progress’);
This opens an overview of the accuracy and loss while training, like this:

I think these two plots/graphs are amazing for the analysis of the CNN. Now, I’ve ran into an issue regarding enabling the option mentioned above.
This plot reporter causes conflicts when trying to use it in a standalone application using MATLAB Compiler. This conflict can be resolved by simply removing the ‘Plots’, ‘training-progress’ from the options. When removing this, you lose the two plots/graphs aswell.
I’ve tried to see where the measurements are stored during the plotting so I could extract the data and make my own plots/graphs. This was without sucess.
Now my question is: Is it possible to make my own accuracy and loss graphs by somehow extracting the data the program would have used to create the plots with the ‘Plots’ option enabled?
Any insight would help.

Greetings,
ThomasWithin the Deep learning Toolbox it’s possible to enable the training plots within the options. It uses the following code:
Options = trainingOptions(‘Plots’, ‘training-progress’);
This opens an overview of the accuracy and loss while training, like this:

I think these two plots/graphs are amazing for the analysis of the CNN. Now, I’ve ran into an issue regarding enabling the option mentioned above.
This plot reporter causes conflicts when trying to use it in a standalone application using MATLAB Compiler. This conflict can be resolved by simply removing the ‘Plots’, ‘training-progress’ from the options. When removing this, you lose the two plots/graphs aswell.
I’ve tried to see where the measurements are stored during the plotting so I could extract the data and make my own plots/graphs. This was without sucess.
Now my question is: Is it possible to make my own accuracy and loss graphs by somehow extracting the data the program would have used to create the plots with the ‘Plots’ option enabled?
Any insight would help.

Greetings,
Thomas Within the Deep learning Toolbox it’s possible to enable the training plots within the options. It uses the following code:
Options = trainingOptions(‘Plots’, ‘training-progress’);
This opens an overview of the accuracy and loss while training, like this:

I think these two plots/graphs are amazing for the analysis of the CNN. Now, I’ve ran into an issue regarding enabling the option mentioned above.
This plot reporter causes conflicts when trying to use it in a standalone application using MATLAB Compiler. This conflict can be resolved by simply removing the ‘Plots’, ‘training-progress’ from the options. When removing this, you lose the two plots/graphs aswell.
I’ve tried to see where the measurements are stored during the plotting so I could extract the data and make my own plots/graphs. This was without sucess.
Now my question is: Is it possible to make my own accuracy and loss graphs by somehow extracting the data the program would have used to create the plots with the ‘Plots’ option enabled?
Any insight would help.

Greetings,
Thomas convolutional neural network, deep learning toolbox, accuracy, loss, plots, extract data, training-progress MATLAB Answers — New Questions

​