Step-by-Step Guide to Running Open Deep Research with smolagents

This guide walks you through setting up and using the Open Deep Research agent framework, inspired by OpenAI’s Deep Research, leveraging Hugging Face’s smolagents library. Follow these steps to reproduce agentic workflows for complex tasks like the GAIA benchmark.

Prerequisites

• Python 3.8+ installed
• Git installed
• Hugging Face Account (optional for some model access)
• Basic familiarity with CLI tools

Step 1: Set Up a Virtual Environment

Create an isolated Python environment to avoid dependency conflicts:

python3 -m venv venv          # Create virtual environment
source venv/bin/activate      # Activate it (Linux/macOS)
# For Windows: venv\Scripts\activate

Step 2: Install Dependencies

1. Upgrade Pip:

   pip install --upgrade pip
   

2. Clone the Repository:

   git clone https://github.com/huggingface/smolagents.git
   cd smolagents/examples/open_deep_research
   

3. Install Requirements:

   pip install -r requirements.txt
   

Step 3: Configure the Agent

Key Components:

• Model: Use Qwen/Qwen2.5-Coder-32B-Instruct (default) or choose from supported models.
• Tools: Built-in tools include web_search, translation, and file/text inspection.
• Imports: Add Python libraries (e.g., pandas, numpy) for code-based agent actions.

Step 4: Run the Agent via CLI

Use the smolagent command to execute tasks:

smolagent "{PROMPT}" \
  --model-type "HfApiModel" \
  --model-id "Qwen/Qwen2.5-Coder-32B-Instruct" \
  --imports "pandas numpy" \
  --tools "web_search translation"

Example: GAIA-Style Task

smolagent "Which fruits in the 2008 painting 'Embroidery from Uzbekistan' were served on the October 1949 breakfast menu of the ocean liner later used in 'The Last Voyage'? List them clockwise from 12 o'clock." \
  --tools "web_search text_inspector"

Model Options

Customize the LLM backend:

Advanced Usage

1. Vision-Enabled Web Browser

For tasks requiring visual analysis (e.g., image-based GAIA questions):

webagent "Analyze the product images on example.com/sale and list prices" \
  --model "LiteLLMModel" \
  --model-id "gpt-4o"

2. Sandboxed Execution

Run untrusted code safely using E2B:

smolagent "{PROMPT}" --sandbox

3. Custom Tools

Add tools from LangChain/Hugging Face Spaces:

# In your Python script
from smolagents import Tool
custom_tool = Tool.from_hub("username/my-custom-tool")

Troubleshooting

Performance Notes

• Code vs. JSON Agents: Code-based agents achieve ~55% accuracy on GAIA validation set vs. 33% for JSON-based (source).
• Speed: Typical response time ~2-5 minutes for complex tasks (varies by model).

Community Contributions

To improve this project:

1. Enhance Tools: Add PDF/Excel support to text_inspector.
2. Optimize Browser: Implement vision-guided navigation.
3. Benchmark: Submit results to GAIA Leaderboard.

By following this guide, you’ve replicated key components of OpenAI’s Deep Research using open-source tools. For updates, star the smolagents repo and join the Hugging Face community! 🚀