Implementation

Refer:

https://www.youtube.com/watch?v=19LQRx78QVU&t=43s

Steps

Install: - anaconda: Jupyter lab is amazing - if error occurs while firing jupyter notebook or lab from cmd: - try: running "activate base" creating an Environment Creation workflows: - install git - create a new env: - go to the folder ImageClassification on cmd - "python -m venv " - Activate the env (in VS code & jupyter lab) - .\imageclassification\Scripts\activate - ( for mac in bash/zsh "source /bin/activate") - pip install ipykernal (didn't work at first, but I installed/update - restart vscode for pip to work in activated venv: ) - check in pip list if installed (ipykernal) - Deactivate: "deactivate" - Use it: - python -m ipykernel install --name=img_class_cnn_venv

• Check how many kernals available:
  • jupyter kernelspec list
• Specifically for vs code:
  • open user settings with ctrl + shift + P
  • type: "env"
  • Python: Create Environment
  • select "Venv"
  • Enter Interpreter path
    • Browse and select python within your virtual env folder
  • Restarting VS code (might not be necessary but)
  • now you see additional kernal with the name of your project

• Specifically for vs code:

  • run jupyter lab or notebook and change the kernel to your virtual env project name

• Uninstalling virtual env:

  • jupyter kernelspec uninstall
  • y

Install TensorFlow for deeplearning

• pip install: https://www.tensorflow.org/install/source_windows
  • Looks at system requirements
  • !pip install tensorflow
  • Diversion: still kernal does not work after selection,
  • Updating the setting jupyter.kernels.trusted:
    • Copy the fully qualified path to the kernelspec, e.g. D:\Py_pro\Image_Class\ImageClassification\imageclassification\share\jupyter\kernels\python3\kernel.json
    • Open the VS User Settings UI using the command Preferences: Open User Settings from the Command Palette.
    • Search for the setting jupyter.kernels.trusted
    • Add a new value into the list using the Add button
      • ex: c:\ProgramData\jupyter\kernels\img_class_cnn_venv\kernel.json
    • Re-load VS Code
    • activate env again
  • Microsoft Visual C++ Build Tools - Desktop dev with C++
    • https://visualstudio.microsoft.com/downloads/

Note: the above steps didn't work, so I tried new steps:

• Create python environment: Venv
• if not automated:
  • activate env
  • go in the env folder (keep it in virtual_envs folder)
  • pip install ipykernal
  • confirm: python -m ipykernel install --name=env_name
  • Open the VS User Settings UI using the command Preferences: Open User Settings from the Command Palette.
    • Search for the setting jupyter.kernels.trusted
    • Add a new value into the list using the Add button
      • ex: c:\ProgramData\jupyter\kernels\img_class_cnn_venv\kernel.json
    • Re-load VS Code
    • activate env again