Installation (Windows)

• git clone https://github.com/silviojaeger/ba_machine_learning_2019
• cd .\ba_machine_learning_2019
• setup.ps1
• Setup VSCode (Optional)
  • Open Folder ba_machine_learning_2019
  • Select Interpreter: command + shift + P execute Python: Select Interpreter select .\env\Scripts\python.exe
  • Activate env (Optional for terminal commands): .\activate.ps1

Upgrade environment

• activate.ps1
• update_requirements.ps1
• python -m pip install <package>
• Commit changes

Create a new environment

• Copy following files to a new foder .<newEnv>:
  • setup.ps1
  • update_requirements.ps1
  • add .\<newEnv>\env to .gitignore

Install graphviz (used to visualize the keras graph)

• Download and install the Windows package from here: https://graphviz.gitlab.io/_pages/Download/Download_windows.html

Install CUDA library(used to train the models on your GPU)

• Update your graphics-card driver: https://www.nvidia.de/Download/index.aspx?lang=de
• Install CUDA 10.0: https://developer.nvidia.com/cuda-10.0-download-archive
• Download cudnn and safe it in C:\tools (you have to create a free NVIDIA-account for this): https://developer.nvidia.com/cudnn
• Add the following paths to the system %PATH% * C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin * C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\extras\CUPTI\libx64 * C:\tools\cuda\bin
• to make it run on a system with cpu's that does'nt support AVX (like the NTB remote-computer) you have to install tensorflow the following:
  • Download this wheel file: https://drive.google.com/open?id=1dWlAqVqcCZmH3q3vefo8ohVhDxBTRp0H
  • Make sure tensorflow is not installed (pip uninstall tensorflow)
  • Install the tensorflow wheele: pip install /PATH-TO-MY-FILE/tensorflow-1.12.0-cp36-cp36m-win_amd64.whl
• More details: https://www.tensorflow.org/install/gpu