This is an End-to-End bare minimum boilerplate Machine Learning application.
Steps...
- Loads a toy dataset
- Performs dummy preprocess step through a data wrangler middleware
- Trains a model and saves it to disk
- Prediction service uses data_wrangler middleware to preprocess and predict the requests received through a Flask REST API layer.
Application contains...
| File | Purpose |
|---|---|
| ML Classifier Sample.ipynb | The classifier, where model is trained |
| data_wrangler.py | Represents middleware to pre-processes data. Consumed by both, training and prediction steps |
| Prediction Service.ipynb | Flask REST API which consumes trained model through data_wrangler |
| iris_clf_model.pkl | Model saved to disk |