hmnshudhmn24/toolplanner-lm

Tool Planner LM 🔧🧠

ToolPlannerLM is an AI planning model that predicts the optimal sequence of tools or APIs an agent should use to complete a given task.

It converts natural-language tasks into structured, explainable execution plans, making it ideal for autonomous agents, workflow automation, and tool-augmented LLM systems.

---

🔍 What Problem Does It Solve?

Agent systems often fail not because they lack intelligence, but because they:

• Choose the wrong tools
• Use tools in the wrong order
• Miss dependencies between actions

ToolPlannerLM solves this by generating a graph-based plan that defines:

• Which tools to use
• In what order
• Why each step is needed

---

✨ Key Features

• 🧠 Natural-language task parsing
• 🔧 Intelligent tool selection
• 🗺️ Graph-based execution planning
• ⚡ Plan optimization (ordering & cost awareness)
• 🧩 Explainable, modular design
• 🤗 Hugging Face–ready (graph-ml pipeline)
• 🎛️ Gradio demo included
• 🧪 Unit-tested core logic

---

📂 Project Structure

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

---

⚙️ Installation

pip install -r requirements.txt

---

🚀 Quick Usage

from src.inference import ToolPlannerPipeline

pipeline = ToolPlannerPipeline()

task = "Find the population of France and calculate growth rate"
result = pipeline(task)

print(result)

---

🎛️ Gradio Demo

python app.py

---

🧠 How It Works

1. Task Parsing
2. Tool Selection
3. Plan Graph Construction
4. Plan Optimization