🧠 Text Summarization App (T5-Base)

An end-to-end Text Summarization Application built using Hugging Face Transformers, FastAPI, and MLflow, and containerized using Docker for seamless deployment on AWS EC2.
This project leverages the T5-base model fine-tuned on a custom dataset to generate concise, human-like summaries for long pieces of text.

---

🚀 Key Highlights

• 🔍 Model Training: Fine-tuned t5-base transformer model using the Hugging Face Trainer API on the SAMSum dataset.
• 📊 Experiment Tracking: Integrated MLflow to log parameters, metrics (ROUGE scores), and model artifacts.
• ⚡ FastAPI Web App: Built an interactive web interface for text summarization with dynamic, word-by-word output.
• 🐳 Dockerized Deployment: Fully containerized the application and deployed it on AWS EC2.
• 🔄 Model Checkpointing: Implemented multiple training checkpoints for resuming interrupted training sessions.
• 🧾 Evaluation Metrics:
  • ROUGE-1: 0.0358
  • ROUGE-2: 0.0000
  • ROUGE-L: 0.0356
  • ROUGE-Lsum: 0.0354

---

🧰 Tech Stack & Tools

Category	Tools/Frameworks
Language	Python
Model	Hugging Face t5-base
Libraries	Transformers, Datasets, Torch, MLflow
Web Framework	FastAPI, Jinja2
Containerization	Docker
Deployment	AWS EC2
Version Control	Git, GitHub
Experiment Tracking	MLflow
Evaluation	ROUGE Metrics

---

🧩 Project Architecture

text-summarizer/ │ ├── src/ │ ├── textSummarizer/ │ │ ├── pipeline/ │ │ │ ├── prediction_pipeline.py │ │ │ └── training_pipeline.py │ │ ├── components/ │ │ ├── entity/ │ │ └── utils/ │ └── main.py │ ├── templates/ │ └── index.html ├── static/ │ └── style.css │ ├── Dockerfile ├── app.py ├── requirements.txt └── README.md

---

💻 How to Run Locally

1️⃣ Clone the Repository

git clone https://github.com/MAbdullah005/Text-Summrizer.git
cd Text-Summrizer
2️⃣ Create a Virtual Environment
bash
Copy code
python -m venv myenv
myenv\Scripts\activate     # On Windows
source myenv/bin/activate  # On Linux/Mac
3️⃣ Install Dependencies
bash
Copy code
pip install -r requirements.txt
4️⃣ Run FastAPI App
bash
Copy code
python app.py
Visit the app in your browser: 👉 http://localhost:8080

🐳 Run with Docker
Build Docker Image
bash
Copy code
docker build -t text-summarization:0.1 .
Run Docker Container
bash
Copy code
docker run -d -p 8080:8080 text-summarization:0.1
App will be available at:
👉 http://0.0.0.0:8080

📦 Docker Hub Image
You can also pull the pre-built image directly from Docker Hub:
🔗 https://hub.docker.com/repository/docker/abdullahali005/text-summarization/general

bash
Copy code
docker pull abdullahali005/text-summarization:latest
🧠 Model Summary
Model Used: t5-base

Framework: Hugging Face Transformers

Training Duration: ~2 hours on GPU 

Evaluation: ROUGE metrics for text summarization

Fine-tuned On: SAMSum dataset (dialogue summarization)

📈 MLflow Tracking Example
All experiments were logged and tracked with MLflow, including:

Training and validation loss

ROUGE metrics

Model artifacts and parameters

Versioned checkpoints for resuming training

🌐 Deployment
Deployed the containerized FastAPI app on AWS EC2, allowing users to summarize large text inputs directly through the browser interface.

🧑‍💻 Author
Muhammad Abdullah Ali
💼 Machine Learning Engineer | MLOps Enthusiast
📍 AWS Certified | Hugging Face | Docker | FastAPI
🔗 GitHub
🔗 Docker Hub

⭐ Acknowledgments
Special thanks to the Hugging Face Transformers team for their incredible open-source tools and to MLflow for providing seamless experiment tracking.