This project shows how to train a model using river online machine learning library river
and log the model using MLflow's custom python function
mlflow.pyfunc.log_model
and import the model in BentoML model store for model serving.
Install requirements with:
pip install -r ./requirements.txt- Train and save model:
python ./train.py- Run the service:
bentoml serveOpen in browser http://0.0.0.0:3000 to predict a value.
curl -X 'POST' 'http://0.0.0.0:3000/predict' \
-H 'accept: application/json' \
-H 'Content-Type: application/json' \
-d '{
"ordinal_date": 836489,
"gallup": 47.843213,
"ipsos": 48.07067899999999,
"morning_consult": 52.318749,
"rasmussen": 50.104692,
"you_gov": 58.636914000000004
}'Sample result:
42.74910074533503
Build Bento using the bentofile.yaml which contains all the configurations required:
bentoml build -f ./bentofile.yamlOnce the Bento is built, containerize it as a Docker image for deployment:
bentoml containerize river_arf_model:latest