Deployments

• https://machine-learning-projects.com/ - EC2 (TypeScript/ Java Spring Boot/ Flask) - NEW

Stack/Technologies

• Java Spring Boot
• AWS EC2
• TypeScript
• Flask
• Pytorch (Python)
• AWS SageMaker/ Lambda (Heroku Version)

PyTorch-Image-Classification-Project

• Personal project involving a Java Spring Boot API, Flask microservice, and TypeScript front end.
• The Java API sends requests to a Flask service hosting a PyTorch fasterrcnn_resnet50_fpn_v2 computer vision model. This is deployed on an EC2 instance with its own Route 53 domain.
• This was originally deployed on Heroku as a TypeScript/Express.js application, using Amazon API Gateway to expose a lambda to send requests to a SageMaker endpoint. I have refactored the project as this endpoint was costing roughly $50 a month.
• Can accept both an image URL or an uploaded image from one's computer. It sends binary data to the lambda/ PyTorch Model either way.

Back End Repo - AWS Endpoint/Model/etc

• https://github.com/Mike11199/PyTorch-Image-Classification/

Video - Old (2023)

• https://www.youtube.com/watch?v=abtdBPFu_yM

EC2 Screenshots

Old Website Screenshots