How To Deploy A Machine Learning Model: Step By Step From Notebook To API Deployment On Render

In this tutorial video, we are working on a #MachineLearningProject with #Python, building a machine learning #API with #FastAPI and deploying the API in the cloud, on #Render (the go-to free hosting platform which replaced #Heroku).

Building a machine learning API, deploying it and have it run in the backend of a web application requires just a little additional effort for you, as a data scientist. But the return of showcasing a product the user can relate to, an application rather than "an abstract ML model" is huge!

What would have a greater impact in the following scenarios: an untagible ML model and some reports or a simple web application that took you an extra day to build, having your ML model run in the back?
working on a start-up idea and ask investors for funding
trying to land you first customer
getting a job as a data scientist
presenting your results to the managers in the company

Contents:
0:00 - Introduction
2:24 - Setting up the Git repository
4:38 - Folder structure of the project
5:37 - Create a Python virtual environment with venv
7:29 - Starting point: the Jupyter Notebook
11:00 - Selecting the code chunks for the API
14:13 - Save a Label Encoder for API
15:45 - API Request and Response with pydantic for data validation
20:16 - API service: where the magic happens
49:09 - Create a machine learning API with FastAPI
54:09 - Send requests to the API with FastAPI docs
57:37 - Send requests to the API with Postman
59:29 - Create a requirements.txt file
1:00:29 - Push code to Github
1:03:48 - Deploy a machine learning API for free

GitHub repository with code and artifacts: https://github.com/giraffa-analytics/...

#datascience #machinelearning #python #ai #artificialintelligence #fastapi #render