This is an example of deploying an SKLearn model using Flask. Endpoints are for both training and prediction. A convenient sheel script is provided to launch the Flask app. You can submit training from command line:
curl -X POST -d '@data/train.json' http://localhost:5000/train
You can make a prediction from command line:
curl -X POST -d '@data/predict.json' http://localhost:5000/predict