mzl-init 1.0.2

A production-ready CLI tool that scaffolds AI/ML projects interactively

🚀 mzl-init

A production-ready AI/ML project scaffolding CLI — like Vite, but for Machine Learning.

Installation

pip install -e .

Usage

Interactive mode (recommended)

mzl-init my_project

Launches an arrow-key prompt to select:

• Project type (Tabular ML / Deep Learning / NLP / Computer Vision / LLM App)
• Framework (PyTorch / TensorFlow / Scikit-learn)
• Optional features: FastAPI, Docker, MLflow, DVC, Hydra

Non-interactive mode (CI/CD)

mzl-init init my_project --type tabular --framework sklearn --docker --mlflow

Check your environment

mzl-init doctor

Version

mzl-init --version

Generated Structure

my_project/
├── configs/            # YAML configuration files
├── data/raw|processed|external
├── docs/
├── models/             # Trained artifacts
├── notebooks/
├── reports/
├── scripts/
├── src/
│   ├── data/loader.py
│   ├── features/build_features.py
│   ├── models/model.py         ← framework-specific
│   ├── training/train.py       ← MLflow-aware
│   ├── evaluation/evaluate.py
│   ├── inference/predict.py
│   └── utils/logger.py
├── tests/
├── deployment/api|docker|cloud
├── .env
├── .gitignore
├── main.py
├── requirements.txt
└── pyproject.toml

License

MIT