Frequently Asked Questions (FAQ)

1. What Python versions are supported?

KiteML requires Python 3.10 or higher (officially supporting Python 3.10, 3.11, 3.12, and 3.13).

2. Why is the PyPI package name kiteml-ai while import is import kiteml?

The distribution package on PyPI is named kiteml-ai to distinguish it on PyPI registry, while the Python code namespace remains import kiteml for developer convenience (identical to scikit-learn -> import sklearn).

3. How does KiteML prevent data leakage?

KiteML's LeakageDetector scans dataset features against target variables to identify near 1.0 mutual information scores and target correlation proxies prior to splitting. Furthermore, transformation parameters are fit exclusively on training folds during cross-validation.

4. What is a .kml package?

A .kml package is a serialized binary archive containing transformation DAG states, feature encoders, scalers, and model weights, secured by a SHA-256 integrity checksum.

5. How do I serve a model as a REST API?

Run kiteml serve model.pkl --port 8000 from your terminal to launch a FastAPI REST server with interactive OpenAPI documentation.