hmnshudhmn24/text-classification-distilbert

⭐ Text Classification Model — DistilBERT

A lightweight and efficient DistilBERT-based text classification model designed for binary text classification tasks such as sentiment analysis, opinion detection, or simple natural-language classification projects.

This repository contains:

• ✔ A clean inference script (model.py)
• ✔ A training script for fine-tuning (train.py)
• ✔ Configuration files (config.json)
• ✔ A model card (model_card.md)
• ✔ Example input samples (example_inputs.txt)
• ✔ requirements.txt for dependencies

---

🚀 Features

• Built on DistilBERT, optimized for speed and smaller memory footprint compared to full BERT.
• Easy to fine-tune on any binary text dataset.
• Works on CPU, GPU, and cloud platforms.
• Minimal and beginner-friendly structure.

---

📂 Repository Structure

.
├── README.md
├── requirements.txt
├── model.py
├── train.py
├── config.json
├── model_card.md
└── example_inputs.txt

---

🛠 Installation

Create a virtual environment (recommended) and install dependencies:

python -m venv venv
source venv/bin/activate   # Unix/macOS
venv\Scripts\activate      # Windows

pip install -r requirements.txt

---

🔍 Inference Example

from model import load_model_and_predict, model_predict

model, tokenizer = load_model_and_predict(load_only=True)
output = model_predict("This movie was fantastic!", model, tokenizer)
print(output)

---

🧠 Model Details

• Model name: text-classification-distilbert
• Base Model: distilbert-base-uncased
• Task: Binary Text Classification
• Language: English
• License: Apache-2.0

---

🧪 Training

python train.py

---

📌 Example Inputs

I really like this!
This is absolutely terrible.

---

⚠️ Limitations

• Binary labels only
• Performance depends on training data
• Possible dataset bias