MLOps Computer Vision project

Simple computer vision project deployed using FastAPI and Docker.

The app offers a detection and a segmentation endpoint.

The Deep learning models used for this projects are YoloV8 models from Ultralytics.

Deploy the whole stack (WebApp | Prometheus | Grafana)

Into the app folder run: docker-compose up --build Prometheus is available at port 9090. Grafana available at port 3000.

App deployment only

Docker usage

• build image: docker build -t container-name .
• run container: docker run -p 8000:port container-name

Custom models

You can deploy your own YoloV8 custom models simply modifying the .pt files in the model.py script and adding them in the models folder.

Author:

Gianmarco Russo