LightGBM Example

This example trains a LightGBM classifier with the iris dataset and logs hyperparameters, metrics, and trained model.

It is modified from MLFlow LightGBM native example: https://github.com/mlflow/mlflow/tree/master/examples/lightgbm/

Running the code

python train.py --colsample-bytree 0.8 --subsample 0.9

You can try experimenting with different parameter values like:

python train.py --learning-rate 0.4 --colsample-bytree 0.7 --subsample 0.8

Then you can open the MLflow UI to track the experiments and compare your runs via:

mlflow ui

Running the code as a project

mlflow run . -P learning_rate=0.2 -P colsample_bytree=0.8 -P subsample=0.9

Serve the model with BentoML

bentoml serve

curl -X POST -H "content-type: application/json" --data "[[5.9, 3, 5.1, 1.8]]" http://127.0.0.1:3000/classify

Build Bento for production deployment

bentoml build

Generate docker image from Bento:

bentoml containerize lgb_iris_service:latest