📰 Fake News Detector using Deep Learning (LSTM)

A deep learning-based Fake News Detection application built using TensorFlow, FastAPI, and Streamlit. The model leverages Natural Language Processing (NLP) techniques with TextVectorization and an LSTM (Long Short-Term Memory) network to classify news articles as Real or Fake.

Note: I wasn't able to get any server hosting service for free which can handle my model loading with all those parameters thats why there is only a local hosting option here!

---

📌 Motivation

This project was built to strengthen my understanding of Deep Learning and Natural Language Processing (NLP) by implementing a complete end-to-end machine learning application.

The primary goals of this project were to:

• Learn text preprocessing using TensorFlow's TextVectorization layer.
• Build and train an LSTM-based neural network for text classification.
• Deploy a trained deep learning model using FastAPI.
• Create an interactive web interface using Streamlit.
• Understand how machine learning models are served in production.

---

🛠️ Tech Stack

Model Training

• TensorFlow
• Pandas
• NumPy

Backend (Model Serving)

• FastAPI

Frontend

• Streamlit

---

📂 Project Workflow

News Article
      │
      ▼
TextVectorization
      │
      ▼
LSTM Model
      │
      ▼
FastAPI Backend
      │
      ▼
Streamlit Frontend

---

🚀 Getting Started

Prerequisites

Note: TensorFlow currently does not support Python 3.14. This project uses Python 3.13, so make sure to create and activate a Python 3.13 virtual environment before installing the dependencies.

---

Installation

1. Clone the repository

git clone https://github.com/SubhamJM/Fake-News-Detector.git
cd Fake-News-Detector

2. Install the dependencies

Windows

pip install -r requirements.txt

Linux/macOS

pip3 install -r requirements.txt

---

Running the Application

Start the FastAPI server

From the project root directory, run:

python api/main.py

or on Linux/macOS:

python3 api/main.py

---

Start the Streamlit frontend

Open another terminal in the project root and run:

streamlit run app/app.py

The Streamlit application will open automatically in your browser.

---

📖 Features

• Fake news classification using an LSTM neural network
• Text preprocessing with TensorFlow TextVectorization
• REST API powered by FastAPI
• Simple and interactive Streamlit interface

---

📷 Preview

• Example of a Fake news:

• Example of a Real news: