mlflow

Project for the Computer Vision course.

Train YOLOv5 for breed classification on Oxford Pets III dataset from scratch on Google Colab, and serve through Dockerized implementation of a flask-based HTML/JS frontend and an asynchronous API service on FastAPI. Use of MLFlow for logging.

MLflow plugin adding OIDC/OAuth 2.1 client authorization

MLflow deployment plugin for Knative targets

MLflow plugins adding AWS RDS IAM auth for tracking and model registry stores

Federated learning project. Using python , pytorch , Flower , Mlflow , Docker Create architecture to create , deploy model with Flower( binary classification) .

Terraform configuration to deploy MLflow in a Kubernetes cluster (using Helm).

The dev environment root module uses Kind to create a local k8s cluster.

this project it to practice all concepts and knowledge in the course mlops-zoomcamp

Een MLFlow wrapper die het makkelijk maakt om samen in dezelfde MLFlow omgeving te werken met bijvoorbeeld Teams. Maakt het daarnaast makkelijker om met standaard machine learning packages (nu alleen nog scikit-learn) modellen en scores te loggen.