Skip to content

Releases: softdevhassan/entify

v4.1.0 — Academic Submission Release

28 Mar 03:56

Choose a tag to compare

Academic Submission Release
ILM College Sargodha | BSCS (24-28) | Group #2
Subject: Artificial Intelligence | Topic: Natural Language Processing


What is Entify?

Entify is a comparative NER engine we built as our AI semester project.
The idea was simple — train a CRF model from scratch and compare it
against spaCy's pre-trained neural model to see how much difference
the architecture actually makes.

This release is the final submitted version after 8 weeks of development.


What it does

  • Runs two NER models side by side — CRF (statistical) and spaCy (neural)
  • CRF model trained on 14,041 sentences from the CoNLL2003 dataset
  • Achieved 90.32% weighted F1 on the CRF side
  • Uses BIO tagging to mark entity boundaries
  • Shows performance comparison in real time
  • REST API included for anyone who wants to integrate it
  • Full docs available: User Guide, Developer Guide, API Reference, Test Labs

Tech Stack

Backend — Flask, sklearn-crfsuite, spaCy (en_core_web_sm)
Frontend — Alpine.js, Tailwind CSS, AOS
Hosting — AWS EC2, Apache, Gunicorn
CI/CD — GitHub Actions, auto deploys to entify.orbin.dev on main push


Quick Start

git clone https://github.com/softdevhassan/entify
cd entify
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
python -m spacy download en_core_web_sm
python app.py


Live Demo

https://entify.orbin.dev


Team

Hassan Ali — Team Lead
Mudassir Ali — Member
Saad Ilyas — Member

Supervised by Sir Abdur-Rehman (Artificial Intelligence)
ILM College Sargodha, affiliated with University of Sargodha