davidstutz/matlab-mnist-two-layer-perceptron

A two layer perceptron implemented in MatLab to recognize handwritten digits based on the MNIST dataset.

matlab-mnist-two-layer-perceptron's Issues

loadMNISTImages function not working on my MATLAB

loadMNISTImages function not working on my MATLAB what do i have to load for this ?

Working

Hiii sir..My name is Ashok. Can you tell me the procedure that how to work with this.. i have cloned the project and in MATLAB which one i have to run to see the output.. Please help me..

wrong mini-batch code

trainStochasticSquaredErrorTwoLayerPerceptron.m line 55
weight is updated at each sample of minibatch, this is not mini-batch. it's online

which .m file is the demo?

which .m file is the demo? or there is no demo in the code

Why mini-batch training not work?

Hi, David.

In the file trainStochasticSquaredErrorTwoLayerPerceptron.m, the weight matrices are updated after propagating and back-propagating on each item, this is the standard SGD optimize, while I try to use mini-batch optimize. I modify inputVector = inputValues(:, n(k)); to inputVector = inputValues(:, n(k): n(k) + minBatch ); and change targetVector to targetVector = targetValues(:, n(k): n(k) + minBatch);. I hope the weight matrices are updated epochs times, and this looks like the mini-batch way, However, the experimental results are totally wrong, correct percent is only 10%.

Would you explain why?
Thanks