How to choose the optimum values of the training options Transfer learning of the pretrain Deep networks in MATLAB

Dear all I have a question, I am using Transfer Learning of the pretraining Deep learning networks I would like to know how to choose the optimum values of the training options(Solver, initial learning rate, max epoch, miniBatchsize etc) should I keep trying different values or choosing the default values, please

Answer 1

[Meet]

Hi Hussain,

To choose optimal training options for transfer learning, start with default settings as a baseline. Use 'adam' or 'sgdm' as solvers, with an initial learning rate around 1e-4 or 1e-3. Set 'maxEpochs' between 5-20, and try mini-batch sizes like 32 or 64. Adjust parameters based on validation performance, employing techniques like learning rate schedules and early stopping to refine results. Experiment with different such parameters to guide your choices.

Hope this clears the doubt!!