Template repository for ML Projects

Setup

• Install poetry

• Install the virtual environment

  poetry install

• Activate the virtual environment

  source $(poetry env info --path)/bin/activate

• Activate pre-commit hooks

  pre-commit install

• Download the raw datasets (optional)

  dvc pull

Current Tasks

• Add functions to create custom pytorch datasets and dataloaders from raw data
• Add tests for these custom datasets and dataloaders
• Add CI pipelines for checking code linting and tests
• Create supervised model for fashion MNIST dataset using pytorch
• Create supervised model for fashion MNIST dataset using pytorch-lightning
• Use dvc to save the models and version them

Tools

• git for code version control
• dvc for data version control
• poetry for python dependency management
• pylint, black for code quality
• pytest for all types of tests
• typer for easy cli
• clearml for experiment tracking and model monitoring

Contributing Instructions

• Fork this repository and make sure that the fork is in-sync before sending a pull request.
• Note all the steps that you think are worth noting along with any good tutorials for the particular task that you are doing.
• Make sure that you are following coding best practices for python.