To add a custom figure to the Experiment Manager's "Built-In Training" feature in MATLAB, you can use the "OutputFcn" option in the training options. This allows you to specify a custom output function that is called once before the start of training, after each training iteration, and once after training has finished. This function has access to information such as the current iteration number, loss, and accuracy. It can be used to perform custom actions during the training process, such as updating plots, logging information, or even stopping training based on specific conditions.
You can refer to this documentation
In the provided example, MATLAB demonstrates the creation of a custom training plot function to display the natural logarithms of gradient norm, step norm, training loss, and validation loss during the training of deep learning neural networks using the "trainnet" function. This custom output function updates a custom plot during training and also has the capability to stop training based on specific conditions.
By using the "OutputFcn" name-value pair argument within the "trainingOptions" function and specifying the custom output function, you can add a custom figure to the "Built-In Training" experiment in the Experiment Manager. The custom figure can then be visualized within the "Review Results" tab within the Experiment Manager interface.
Hope this helps!
