News Classifier As Service (2020)
Author: Winnie Yeung
Deploy NLP model for news classification using Flask, Docker, Kubernetes
Goal:
- Deploy ML learning model for inferencing using Docker and Kubernetes
How to use:
- Train NLP model using scikit-learn baseline
- Package environement and dependencies in Docker
- Deploy using Kubernetes (Ingress, Serving)
- Use API to send text in, preprocess and output prediction
Kubernetes Terminology:
- Ingress: External load balancer
- Serving: Internal load balancer
- Pod : Container
Data:
- BBC News Article - 5 News Category classfication
- kaggle datasets download -d yufengdev/bbc-fulltext-and-category
Resources: