Streamlit as a Model Serving Endpoint

Model is trained via train.py and a pickle file was generated in /model/boston_price.pkl To run app to score model in real time via user inputs:

conda create -name streamBoston -file requirements.txt (this installs streamlit and other python package dependencies)
conda activate streamBoston
Run streamlit run app/app.py

This would give you a local URL to access, mine was http://localhost:8501

Here is a screenshot of the web app:

gityow/streamlit_app