Auxiliary future state generation framework using diffusion model

---

📖 Table of Contents

• 🖼 Architecture
• ✨ Features
• ⚙️ Requirements
• 🚦 Quick Start
• 🤝 Contributing
• 📄 License
• 📚 Citation

---

🖼 Architecture

---

⚙️ Requirements

Install dependencies:

1. use one conda env to install habitat-sim
2. use another one conda env to install unik3d

---

🚦 Quick Start

cd axground3d

# use habitat-sim env
python ./run_gen_real.py

# use unik3d env
python ./run_gen_ply.py

# use habitat-sim env
python ./run_gen_fake.py

# Evaluation
python ./fid_cal.py

---

🛠 Configuration

All hyperparameters live in .default.yaml. Make one .local.yaml first. All hyperparameters live in .local.yaml. Key sections:

mp3d_habitat_scene_dataset_path: "<your path>/mp3d/"
vlm_model_path: "./model/Janus-Pro-1B"
number_of_node_per_scene: 1

---

🤝 Contributing

1. Fork → Clone → Create feature branch
2. Add tests for new modules
3. Submit PR → Review → Merge

---

📄 License

This project is MIT Licensed. See LICENSE for details.

---

📚 Citation

If you find our work helpful, feel free to give us a cite:

@misc{axground3D,
    title = {Auxiliary future state generation framework using diffusion model},
    url = {https://github.com/YichengDuan/axground3D},
    author = {Yicheng Duan, Duo Chen},
    month = {April},
    year = {2025}
}