hmnshudhmn24/intentgraph-lm

IntentGraphLM 🚀

IntentGraphLM is an open-source language model pipeline that converts raw natural language into dynamic intent graphs, enabling smarter agents, workflow orchestration, and multi-step task planning.

This project is designed to be lightweight, extensible, and Hugging Face–ready, with strong future scope in agentic AI systems.

---

🔍 What Problem Does It Solve?

Traditional intent detection returns flat labels.
IntentGraphLM goes further by:

• Extracting multiple intents
• Modeling relationships between intents
• Representing them as a directed graph
• Enabling dynamic updates during conversations

---

✨ Key Features

• 🧠 Intent extraction from raw text
• 🕸️ Graph-based intent representation
• 🔁 Dynamic intent graph updates
• ⚡ Lightweight, dependency-minimal design
• 🤗 Hugging Face–compatible inference pipeline
• 🧩 Modular and extensible architecture

---

📂 Project Structure

intentgraph-lm/
├── config/
├── data/
├── src/
├── training/
├── scripts/
├── tests/
├── notebooks/
├── README.md
├── model_card.md
├── requirements.txt
└── LICENSE

---

⚙️ Installation

pip install -r requirements.txt

---

🚀 Quick Usage

from src import IntentGraphPipeline

pipeline = IntentGraphPipeline()

output = pipeline("Book a flight and then reserve a hotel")

print(output)

---

🧠 How It Works

1. Tokenization
2. Intent Extraction
3. Graph Building
4. Dynamic Updates

---

🧪 Testing

pytest tests/