Deep Learning workshop workspace

This repo contains datasets and Machine Learning tasks/solutions related the the Machine Learning workshop.

Getting started

• Register on GitHup if not yet there
• Fork this repository (we will work via pull requests)
• Create your personal branch in forked repo (name is up to you)
• Install python (2.7 is OK)
• Install required python modules:

    # numpy to be installed (for vectorized computations)
    sudo pip install numpy

    # imeges manipulation library
    sudo pip install imageio

    # matplotlib to be installed (for plotting the data)
    sudo pip install matplotlib

    # h5py to be installed (for H5 files support,
    # which are designed to store nonuniform
    # data structures, like various data-sets)
    sudo pip install h5py

    # [optional] pyinstaller to be installed (for standalone
    # python scripts executables)
    # NOTE: you don't install pyinstaller, then
    #	do not set GENERATE_PYTHON_STANDALONE_EXECUTABLES
    #   CMake parameter to True.
    sudo pip install pyinstaller

General workflow

1. Commit your solution to correct branch in forked repo,
   • if you need to commit new data to the dataset, use master branch to make your examples be available to everyone.
   • if you need to commit your solution for the task
     • use your personal branch
     • commit to subfolder of relevant workshop folder: e.g. for John Smith:

        <branch jsmith>
        PROJECT_ROOT/workshop2/jsmith/mycoolsolution1.py
        PROJECT_ROOT/workshop2/jsmith/mycoolsolution2.py
       

   • if you need to commit some reasonable improvement for the whole project, please also use master branch
2. Create pull request for common repository.
3. Update your commit according to comments.
4. As your pull request is approved, your changed will be added to the main repo.

For convenience

You can add custom run procedures in your IDE, to make it convenient to start relevant python script directly from IDE.