Add demo dir and commands

demo/demo_commands.md

@@ -0,0 +1,88 @@
+# MLflow demo using project examples
+
+This demo will use the example in the `mlflow/examples/sklearn_elasticnet_wine` directory, reacheable as `../examples/sklearn_elasticnet_wine`.
+
+Run all the commands below from the `demo` directory.
+
+## 1) Launch MLflow UI
+
+```
+mlflow ui
+```
+
+## 2) MLflow Tracking
+### Run once logging random params & metrics values
+
+```
+python ../examples/quickstart/mlflow_tracking.py
+```
+
+### Run multiple times to show UI & graphs
+
+```
+for i in {1..100}; do python ../examples/quickstart/mlflow_tracking.py; done
+```
+
+
+# 3) MLflow projects - Running the training & tracking parameters
+
+Running this command will fail if you don't have all the lis and dependencies installed:
+
+```
+python ../examples/sklearn_elasticnet_wine/train.py
+```
+
+MLflow can run it for you and install dependencies...
+
+```
+mlflow run ../examples/sklearn_elasticnet_wine  // will tell you the parameters needed to run
+```
+
+```
+mlflow run ../examples/sklearn_elasticnet_wine -P alpha=0.4
+```
+
+```
+mlflow run ../examples/sklearn_elasticnet_wine -P alpha=0.4 -P l1_ratio=0.12
+```
+
+
+## Let's run it a few times...
+```
+./multiple_runs.sh
+```
+
+## Run projects from a git repo
+
+```
+mlflow run https://github.com/mlflow/mlflow-example.git -P alpha=0.5
+```
+
+```
+mlflow run --experiment-name=sklearn_elasticnet_wine https://github.com/mlflow/mlflow-example.git -P alpha=0.5
+```
+
+# Serving the model and making predictions
+
+## Serving the model locally
+
+Look for the `model_file_path` in the UI, then tun this command to serve the model on port 1234:
+
+```
+mlflow models serve -m {model_file_path} -p 1234
+```
+
+## Predict
+
+To make a prediction using the model served on port 1234, tun:
+
+```
+curl -X POST -H "Content-Type:application/json; format=pandas-split" --data '{"columns":["alcohol", "chlorides", "citric acid", "density", "fixed acidity", "free sulfur dioxide", "pH", "residual sugar", "sulphates", "total sulfur dioxide", "volatile acidity"],"data":[[12.8, 0.029, 0.48, 0.98, 6.2, 29, 3.33, 1.2, 0.39, 75, 0.66]]}' http://127.0.0.1:1234/invocations
+```
+# Cleanup
+
+```
+rm -r mlruns
+rm -r outputs
+```
+

demo/multiple_runs.sh

@@ -0,0 +1,22 @@
+#!/bin/bash
+
+alpha=(0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8)
+l1=(0.1 0.2 0.3)
+
+experiment_dir=$1
+experiment_name="$(basename "$experiment_dir")"
+
+echo "running experiment in: $experiment_dir"
+echo "experiment name: $experiment_name"
+
+# Create experiment if it doesn't already exist
+if ! (mlflow experiments list | grep $experiment_name); then
+	mlflow experiments create -n $experiment_name
+fi
+
+# Run the experiment a few times changing the parameters
+for x in "${alpha[@]}"; do
+	for y in "${l1[@]}"; do
+		mlflow run --experiment-name="$experiment_name" "$experiment_dir"  -P alpha=$x -P l1_ratio=$y
+	done
+done