AlyaAloft: Advanced Document Q&A

Overview

AlyaAloft is a powerful document question-answering application that uses the FLAN-T5 language model with enhanced prompting techniques to provide high-quality responses to user queries about document content.

Key Features

• Advanced Prompting Techniques: Domain-specific templates, chain-of-thought reasoning, and iterative refinement for complex questions
• Optimized Model Performance: 8-bit quantization for CUDA-enabled devices to reduce memory usage while maintaining quality
• PDF Document Processing: Extract and chunk document content for efficient retrieval
• Conversation Memory: Maintain context across multiple user queries
• Responsive Web Interface: Clean, modern UI for document upload and querying

Screenshots

Demo & Documentation

• 📄 Project Documentation
• 🎥 Video Preview

Getting Started

Prerequisites

• Python 3.8+ with pip
• PyTorch with CUDA support (recommended for faster inference)
• 4GB+ RAM (8GB+ recommended)
• 2GB+ free disk space for models

Installation

1. Clone the repository:

   git clone https://github.com/Prathameshv07/AlyaAloft.git
   cd AlyaAloft

2. Create a virtual environment in python or conda:

   # create a python virtual environment
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   
   # create a conda virtual environment
   conda create -n venv python=3.9
   conda activate venv

3. Install dependencies:

   pip install -r requirements.txt

4. Download the FLAN-T5 model:

   python scripts/download_t5_model.py

   For a smaller model (better for limited resources):

   python scripts/download_t5_model.py --model google/flan-t5-small

   For better quality (requires more RAM):

   python scripts/download_t5_model.py --model google/flan-t5-large

Running the Application

Start the application with:

python start_app.py

By default, the server will run on http://127.0.0.1:8000.

Command-line options:

python start_app.py --host 0.0.0.0 --port 9000 --log-level DEBUG

Usage

1. Open the web interface in your browser at http://127.0.0.1:8000/chat
2. Upload a PDF document using the upload button
3. Ask questions about the document in natural language
4. View responses with reference to the source document

License

This project is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License.
You are free to use, share, and adapt the material for non-commercial and educational purposes, as long as proper credit is given and any changes are noted.

Learn more: http://creativecommons.org/licenses/by-nc/4.0/