# PartPacker.net - AI Model Information

## Website Overview
PartPacker.net is the official information hub for NVIDIA's PartPacker technology - a revolutionary AI system that transforms single 2D images into editable 3D models with separable parts.

## Key Information

### Technology
- **Name**: PartPacker
- **Developer**: NVIDIA Research Labs
- **Type**: 3D Generation AI Model
- **Architecture**: Diffusion Transformer (DiT)
- **Innovation**: Dual-volume packing strategy for part-based 3D generation

### Capabilities
- Single 2D image input (optimal: 518×518 pixels)
- Generates interconnected 3D parts
- Editable and separable components
- Multiple export formats (STL, GLTF, OBJ, 3MF)
- Suitable for 3D printing, animation, and content creation

### Technical Requirements
- GPU: 16GB+ VRAM recommended
- CUDA-compatible NVIDIA GPU
- Python 3.8+, PyTorch
- Supports JPEG, PNG, BMP image formats

### Use Cases
1. **Content Creation**: Game assets, film props, VR content
2. **3D Printing**: Rapid prototyping, multi-material printing
3. **Education & Research**: Academic research, algorithm development
4. **Product Design**: Concept visualization, design validation

### Licensing
- Available for non-commercial use
- Commercial licensing available through NVIDIA Research
- Open-source code available on GitHub

### Resources
- **GitHub**: https://github.com/NVlabs/PartPacker
- **Hugging Face**: https://huggingface.co/nvidia/PartPacker
- **Research Paper**: https://research.nvidia.com/labs/dir/partpacker/
- **Demo**: https://nvidia-partpacker.hf.space

### FAQ Highlights
- Processing time: 2-5 minutes on recommended hardware
- Individual part editing supported
- Multiple output formats available
- Automatic image resizing to optimal resolution

## Contact & Support
For technical questions, commercial licensing, or research collaboration, contact NVIDIA Research Labs.

---
Last Updated: December 2024
Website: https://partpacker.net