A lightweight FastAPI application for serving machine learning predictions via HTTP.

• Clone https://github.com/amitesh786/FastAPI-ML or create your project directory

  • cd FastAPI-ML

• Create and activate a virtual environment

  • python3 -m venv fastml
  • source fastml/bin/activate

• Install dependencies

  • pip install numpy==1.23.5 scikit-learn==1.1.3 pandas fastapi uvicorn

• Installing from requirements.txt

  • pip install -r requirements.txt

• Dependencies section

  • Let users install all dependencies in one go: pip freeze > requirements.txt

• python -m uvicorn mlapi:app --reload

  • Ping GET: http://127.0.0.1:8000/ping - Checks if the API is live.
    • Response : { "message": "API is live!" }
  • POST /: http://127.0.0.1:8000 - Submit data for prediction.
    • Request Body (JSON):
      • { "YearsAtCompany": 5, "EmployeeSatisfaction": 1, "Position": "Manager", "Salary": 2.0 }
    • Response : { "prediction": 1 }

• Example request (using curl or Postman):
  • curl -X POST "http://127.0.0.1:8000/" -H "Content-Type: application/json" -d '{ "YearsAtCompany": 3, "EmployeeSatisfaction": 4.2, "Position": "Manager", "Salary": 2 }'

• ModuleNotFoundError: Ensure your virtualenv is activated before running anything.
• Binary Incompatibility Errors (numpy.dtype size changed):
  • Likely due to version mismatch between model training and current environment.
  • Recreate environment using exact versions.

• mlapi.py – FastAPI app with prediction endpoint.
• rfmodel.pkl – Pre-trained ML model.

• Make sure rfmodel.pkl was trained using the same scikit-learn and numpy versions.
• If you face issues like numpy dtype size changed, recreate the virtualenv with matching versions.

• Clone the repository: git clone https://github.com/amitesh786/FastAPI-ML

• Feel free to contribute or suggest improvements!