VOLMO

Versatile and Open Large Models for Ophthalmology
Official Repository

---

📋 Overview

Official repository of VOLMO: Versatile and Open Large Models for Ophthalmology.

✨ Key Features

• 🤖 Models - Pre-trained VOLMO models for ophthalmologic vision-language tasks
• 📊 Data for Evaluation - Evaluation datasets for model assessment

---

🤖 Models

VOLMO-2B is a 2-billion parameter mulimodal large language model. The pre-trained model weights are available on Hugging Face.

Download Model

from huggingface_hub import snapshot_download

# Download VOLMO-2B model
model_path = snapshot_download(
    repo_id="Yale-BIDS-Chen/VOLMO-2B",
    local_dir="./models/volmo_2b_stage_3"
)

Hugging Face Repository: Yale-BIDS-Chen/VOLMO-2B

For detailed model documentation, see models/README.md.

---

🚀 Quick Start

Environment Setup

# Create conda environment
conda create -n volmo python=3.9
conda activate volmo

# Install requirements
pip install -r requirements.txt

Run Complete Evaluation

python launch.py

The pipeline will automatically:

• ✅ Load VOLMO model
• ✅ Process all 4 task categories: binary classification, staging, image description, and clinical assessment
• ✅ Generate timestamped report: volmo_evaluation_results/report_YYYYMMDDHHMMSS.md

---

📄 Output Reports

Reports are generated in volmo_evaluation_results/ with the following structure:

# VOLMO Evaluation Report
**Generated:** 2026-01-10 16:45:23

## Binary Classification
**Status:** 9/9 datasets completed

### BRSET AMD
- **Accuracy:** 0.9259
- **Samples:** 108

### BRSET DR
- **Accuracy:** 0.8914
- **Samples:** 442
...

---

📝 License

This project is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0).

---

🤝 Contributing

Contributions are welcome! Please feel free to submit issues or pull requests.

---

⚠️ Disclaimer

The information produced on this website is not intended for direct diagnostic use or medical decision-making without review and oversight by a clinical professional. Individuals should not change their health behavior solely on the basis of information produced on this website

---

📧 Contact

For questions or collaboration inquiries, please email:

• Zhenyue Qin: kf.zy.qin@gmail.com
• Qingyu Chen: qingyu.chen@yale.edu

---

Built with ❤️ for exploring AI's potential in ophthalmology

VOLMO: Versatile and Open Large Models for Ophthalmology