ML Blueprint

A starter template for machine learning projects as part of my newsletter series: https://www.sarahglasmacher.com/ml-repo-structure-challenge/

Features

• Project structure for ML workflows using a Python package for easier deployment later on
• Example scripts for data processing, training, and evaluation
• Using Pydantic for safe and verified configuration .yamls
• Pyproject.toml file for dependency and project management

Getting Started

For the repository to make it your own, then:

1. Clone the repository:

   git clone https://github.com/yourusername/ml-blueprint.git
   cd ml-blueprint

2. Install dependencies:

   uv sync

Example Data

Uses example data from the following Kaggle dataset for demonstration purposes only: https://www.kaggle.com/competitions/playground-series-s4e12/data

To use the repo, add the data yourself in the following structure:

Project Structure

ml-blueprint/
.
├── artifacts/
...
├── data                            # add here
│   ├── playground-series-s4e12     # add this repository by downloading from kaggle
│   │   ├── sample_submission.csv
│   │   ├── test.csv
│   │   └── train.csv
│   ├── test.csv                    # these will be generated by code
│   ├── train.csv
├── notebooks
│   └── one_stop_notebook.ipynb
├── scripts/
...
├── src/
│   └── <your_pkg>/
...
├── pyproject.toml
├── README.md
└── uv.lock