Skip to content

Latest commit

 

History

History
30 lines (23 loc) · 1.02 KB

README.md

File metadata and controls

30 lines (23 loc) · 1.02 KB

HuggingFace-URL-classification

Description

nlp_classification.ipynb: Exploring Machine Learning (ML) models accuracy for text classification through systematic experiments with varying features (title, content, URL) and models (8 traditional ML models, HuggingFace Transformers). Achieved up to 95% accuracy, revealing key insights on optimal input-model combos

Deployment: Deploy the Random Forest utilising FastAPI and Docker

Getting Started

Step-by-step instructions for setting up and running the app

Setting Up

git clone https://github.com/DimitrisParaskevopoulos/HuggingFace-text-classification.git
cd your-local-path/HuggingFace-text-classification

Running

docker-compose up --build -d 

Testing

curl --location "http://localhost:5000/predict/rf" ^
--header "Content-Type: application/json" ^
--data "{""url"": ""https://www.sport24.gr/football/real-mpartselona-3-2-o-vasilias-mpeligcham-milise-sto-91-kai-estepse-protathlites-toys-merengkes.10302657.html""}"