README.md

metadata

title: DetectAI
emoji: 🕵️
colorFrom: gray
colorTo: blue
sdk: docker
pinned: false

DetectAI

A local, offline AI text detection tool powered by RoBERTa and GPU Acceleration.

🚀 Features

• Offline Analysis: Uses a local Transformer model (roberta-base) to detect AI-generated text.
• GPU Accelerated: Optimized for NVIDIA GPUs (CUDA 12.1) for <1s inference.
• Visual Reports: Downloadable PDF reports with AI-highlighted text.
• Modern UI: Built with React + Vite, featuring a glassmorphism design.

🛠️ Tech Stack

• Backend: Python, Django, Django Rest Framework
• AI Engine: PyTorch, Transformers (Hugging Face)
• Frontend: React, TailwindCSS, Framer Motion
• PDF Generation: ReportLab

📦 Installation

Prerequisites

1. Python 3.10+
2. Node.js & npm
3. NVIDIA GPU Driver (Version >= 531.x for CUDA 12.1 support)

1. Backend Setup

cd backend
python -m venv venv
# Activate venv:
# Windows: .\venv\Scripts\activate
# Linux/Mac: source venv/bin/activate

# Install Dependencies (includes PyTorch CUDA)
pip install -r requirements.txt
# OR manually:
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
pip install django djangorestframework transformers reportlab

2. Frontend Setup

cd frontend
npm install

🏃‍♂️ Running the App

Step 1: Start Backend (Port 8000)

cd backend
python manage.py runserver

Step 2: Start Frontend (Port 5173)

cd frontend
npm run dev

Step 3: Open App

Go to http://localhost:5173 in your browser.

⚡ Performance Tuning

The backend is pre-optimized for 4GB VRAM cards (like GTX 1650):

• Batch Size: 8
• Strided Masking: Analyzes 33% of tokens for 3x speedup. To change these, edit backend/detector/views.py.

🚀 Deployment (Hugging Face Spaces)

This project relies on Docker and GitHub Actions for deployment.

1. Prerequisites

• A Hugging Face Account.
• A Private Space created on Hugging Face (Select "Docker" as the SDK).

2. Configure GitHub Secrets

Go to your GitHub Repository -> Settings -> Secrets and variables -> Actions -> New repository secret.

Add the following secrets:

Secret Name	Value
HF_TOKEN	Your Hugging Face Access Token (Write permissions).
HF_USERNAME	Your Hugging Face Username (e.g., vivek1192).

3. Trigger Deployment

1. Commit your changes:

   git add .
   git commit -m "Setup CI/CD for Hugging Face"
   

2. Push to develop branch: (The workflow is configured to trigger on develop)

   git checkout develop
   # Merge your changes if needed
   git push origin develop
   

The GitHub Action will:

1. Build the Docker image (React Frontend + Django Backend).
2. Push it to your Hugging Face Space.