Add Validation to training

Add validation to runtraining

PreprocessValidation.py ~/validationimages ~/validationlabels ~/validationdata
PreprocessTrainingData.py ~/trainingimages ~/traininglabels ~/AugTrData
runtraining.sh --validation_dir ~/validationdata/ --numiterations 50000 ~/AugTrData ~/trained_network

Plot loss and accuracy

Run PlotValidation on the log folder of each model

PlotValidation.m ~/trained_network/1fm/log

This will create plots of loss and accuracy and save them as .pdf in the same log folder.