Folder Structure for Training Data in Encrypted Traffic Classification Task

[finleyb]

For the encrypted traffic classification task, the Dataset.txt and Png2Mnist.py files seem to imply that each class (label) should have it's own folder with associated pcap files inside (in other words the label information is determined by the folder structure). However the Pcap2Session and ProcessSession files seem to assume all pcap files are together in a single folder (for example gci just looks within the single folder).

Maybe I am missing something about these assumptions?