NLP-sequence-classification

BESSTIE: Sentiment & Sarcasm Classification across English Varieties University of Surrey ,Semester 2, 2026

Setup

git clone https://github.com/momofahmi/NLP-sequence-classification.git
cd NLP-SEQUENCE-CLASSIFICATION
pip install -r requirements.txt

How to get the dataset in your local folder:

from datasets import load_dataset
ds = load_dataset("surrey-nlp/BESSTIE-CW-26")

A .env file at the project root is required for HuggingFace authentication.

main.ipynb must stay at the repository root.

---

Deployed app

The Gradio app combines all 5 models (3 LoRA adapters for sarcasm + 2 RoBERTa models for sarcasm and sentiment) and is publicly accessible at:

https://huggingface.co/spaces/momofahmi/besstie-cw-nlp

The app has two tabs: a single-text mode where the user explicitly selects an English variety (per the coursework spec), and a comparison mode that runs every text through all 5 models in parallel.

---