Reproducibility&Workflow

[Guldanika]

Course

machine-learning-zoomcamp

Question

Are there common pitfalls to watch for when combining notebooks, API code, and Docker for reproducibility in larger projects?

I successfully deployed the model with Docker. Are there common pitfalls to watch for when combining notebooks, API code, and Docker for reproducibility in larger projects?

Answer

Yes, Docker helps a lot with reproducibility, but common pitfalls include missing or mismatched dependencies, hardcoded data paths, and version differences between environments. It’s also important to save trained models and test the API endpoints to ensure everything works together.

Code example(Dockerfile)
FROM python:3.13.5-slim-bookworm
WORKDIR /code

Install dependencies

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

Copy application code and model

COPY app/ ./app/

Run FastAPI app

CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "8000"]

Checklist

• I have searched existing FAQs and this question is not already answered
• The answer provides accurate, helpful information
• I have included any relevant code examples or links

[github-actions]

❌ FAQ Bot encountered an error processing this proposal.

Please check the workflow run for details.

A maintainer will review this manually.