Skip to content

A starter project template with TensorFlow 2.9 on Python 3.8+

License

Notifications You must be signed in to change notification settings

yashgorana/tensorflow-starter-project

Repository files navigation

Tensorflow (2.9) starter project

A Data Science starter project with Tensorflow 2.9 on Python 3.8+.

Setup

To keep things tidy, this project uses poetry for managing dependencies inside a virtual environment.

Run the following to install poetry

$ pip install poetry

If you're on Windows, some dependency installations might break due to Windows long path limitation. You must run assets/enable-win-long-paths.reg and reboot your system to fix this issue.

Now simply run the following to install the standard data science stack

$ poetry install

Poetry uses the following files:

  • pyproject.toml contains all dependencies
  • poetry.toml contains the Poetry configuration. By default the project has configured the virtual environment to be setup inside this project itself.
  • poetry.lock file that contains the file signatures for packages and ensures that dependency tree is frozen, identical for all installs and hence reproducable.

Project Structure

Provisions for Visual Studio Code

The project includes a .vscode directory that will configure your VS Code workspace to support Python development.

Workspace Settings (settings.json)

For now, the workspace is configured to format code using black

Debug Settings (launch.json)

A debug launch configuration has been provided that will debug the open file. To avoid any sys.path hacks for sibling imports, it by default injects current working directory as PYTHONPATH.

Extensions (extensions.json)

The following extensions will be recommended by VS Code

  • Python
  • Pylance (Python language servers)
  • Jupyter & Jupyter Renderers

License

The source code for the site is licensed under the MIT license, which you can find in the LICENSE file.