# Example MLflow project

## Overview

This is an example MLflow project for the [MLflow Quickstart](https://www.mlflow.org/docs/latest/quickstart.html#quickstart)
documentation.

Using both the [UCI Wine Quality dataset](http://archive.ics.uci.edu/ml/datasets/Wine+Quality) (by P. Cortez, A. Cerdeira, F. 
Almeida, T. Matos and J. Reis.) and Elastic Net to predict quality, we create an MLflow project.

Moreover, The example uses MLproject to set up a Conda environment, define parameter types and defaults, entry point for training, etc.

## Instructions

If you reached this repository from going through the [MLflow Quickstart](https://www.mlflow.org/docs/latest/quickstart.html#quickstart)
documentation, please follow these instructions (else, feel free to independently go through the material yourself):

1 ) Per the [Running MLflow Projects](https://www.mlflow.org/docs/latest/quickstart.html#running-mlflow-projects) section, run the following:

`mlflow run tutorial -P alpha=0.5`

`mlflow run git@github.com:mlflow/mlflow-example.git -P alpha=5`

> **Note:** If you are receiving git permission issues, please _git clone_ this repository. 
>
> Thereafter, run command: 
>
> `mlflow run mlflow-example -P alpha=0.5`

2 ) Per the [Saving and Serving Models](https://www.mlflow.org/docs/latest/quickstart.html#saving-and-serving-models) 
section, run:

`python sklearn_logistic_regression/train.py`

Therafter, to serve the scikit-learn model through a REST server, run:

`mlflow sklearn serve -r <RUN_ID> model`

Now, run 

`curl -d '[{"x": 1}, {"x": -1}]' -H 'Content-Type: application/json' -X POST localhost:5000/invocations`


Congrats, you are done! Please refer back to the [MLflow Quickstart](https://www.mlflow.org/docs/latest/quickstart.html#quickstart) documentation or feel free to play around more with MLflow!

Cheers!