RDAWG 3D Pack - CUDA 12.8 + PyTorch 2.9.0 Optimized

🔷 Modern 3D Processing Nodes for ComfyUI

A comprehensive, next-generation 3D processing toolkit specifically optimized for CUDA 12.8 and PyTorch 2.9.0. This custom node package provides advanced 3D model loading, processing, transformation, and rendering capabilities with GPU acceleration.

✨ Key Features

🚀 Performance Optimized

• CUDA 12.8 Native Support - Leverages latest CUDA features
• PyTorch 2.9.0+ Compatible - Built for modern PyTorch architecture
• GPU Memory Efficient - Smart memory management for large 3D models
• Batch Processing Support - Process multiple 3D models simultaneously

🔧 Core Functionality

• 3D Model Loading - Support for OBJ, STL, PLY, GLTF formats
• Mesh Processing - Vertex/face manipulation, subdivision, smoothing
• Point Cloud Operations - Filtering, downsampling, clustering
• 3D Transformations - Scale, rotate, translate with precision control
• Neural Rendering - AI-powered 3D to 2D conversion
• File Format Conversion - Comprehensive 3D format support

🎯 Advanced Features

• GPU-Accelerated Rendering - Real-time 3D visualization
• Custom Shader Support - GLSL shader integration
• Batch Operations - Process multiple meshes in parallel
• Memory Management - Automatic VRAM optimization
• Error Recovery - Robust error handling and fallbacks

📦 Installation

⚠️ Important: Install PyTorch CUDA First!

RDAWG 3D Pack requires PyTorch with CUDA support for 3D processing:

# Install PyTorch 2.9.0 with CUDA 12.8 support
pip install torch==2.9.0+cu128 torchvision==0.24.0+cu128 torchaudio==2.9.0+cu128 --index-url https://download.pytorch.org/whl/cu128

🚀 Recommended Installation Method

cd ComfyUI/custom_nodes
git clone https://github.com/rdawgemfl/rdawg-3d-pack-cu128-pytorch-2.9.0
cd rdawg-3d-pack-cu128-pytorch-2.9.0

# Run the automated installer (installs CUDA PyTorch + dependencies)
python install.py

Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/rdawgemfl/rdawg-3d-pack-cu128-pytorch-2.9.0
cd rdawg-3d-pack-cu128-pytorch-2.9.0

# 1. Install PyTorch CUDA first (required!)
pip install torch==2.9.0+cu128 torchvision==0.24.0+cu128 torchaudio==2.9.0+cu128 --index-url https://download.pytorch.org/whl/cu128

# 2. Install the package
pip install -e .

Optional Dependencies

# Full 3D processing suite
pip install -e .[full]

# Neural rendering capabilities
pip install -e .[neural]

**Note**: This option will build xformers from source to ensure CUDA 12.8 compatibility with PyTorch 2.9.0. The build process can take 10-30 minutes and requires proper CUDA 12.8 installation.

# 3D reconstruction tools
pip install -e .[reconstruction]

Open3D Installation (Recommended)

For the best 3D processing experience, install Open3D 0.19.0 manually:

🚀 Automatic Download:

python download_open3d.py

📦 Manual Download: Download wheels from: https://github.com/isl-org/Open3D/releases/tag/v0.19.0

Windows Installation:

# Python 3.12
pip install open3d-0.19.0-cp312-cp312-win_amd64.whl

# Python 3.11
pip install open3d-0.19.0-cp311-cp311-win_amd64.whl

# Python 3.10
pip install open3d-0.19.0-cp310-cp310-win_amd64.whl

🎮 Usage

🆕 Important: 3D Models Required

RDAWG 3D Pack works with 3D geometric models (.obj, .stl, .ply files), NOT AI models!

📚 Quick Start:

• Built-in test models: Run python create_test_models.py to generate test_cube.obj, test_sphere.obj, test_pyramid.obj
• Download models: See docs/3D_MODELS_GUIDE.md for where to download free 3D models
• Supported formats: OBJ, STL, PLY, GLTF files

🌐 Where to Get 3D Models:

• Sketchfab (free section) - High quality models
• Thingiverse - 3D printable models
• Free3D - Various categories
• CGTrader - Mixed free/paid models

Basic 3D Model Loading

1. Add 🔷 Load 3D Model (RDAWG+Open3D) node
2. Specify your 3D file path (e.g., test_cube.obj or downloaded models)
3. Configure loading options (texture, normalization, device)
4. Connect to downstream processing nodes

3D Transformations

1. Load your 3D model
2. Add 🔷 Transform 3D Mesh (RDAWG) node
3. Adjust scale, rotation, and translation parameters
4. Connect transformed mesh to rendering or export nodes

3D to 2D Rendering

1. Load or create a 3D mesh
2. Add 🔷 3D to Image (RDAWG) node
3. Configure rendering parameters (resolution, lighting, camera)
4. Output rendered image to standard ComfyUI image nodes

🔧 Node Reference

Core Nodes

| Node | Description | Inputs | Outputs | |------|-------------|--------|---------| | 🔷 Load 3D Model | Load 3D models from files | file_path, options | mesh, info | | 🔷 Create 3D Mesh | Create mesh from tensors | vertices, faces | mesh, info | | 🔷 Transform 3D Mesh | Apply transformations | mesh, transform params | transformed_mesh | | 🔷 3D to Image | Render 3D to 2D | mesh, render settings | image |

Advanced Nodes (Optional Dependencies)

| Node | Category | Description | |------|----------|-------------| | Point Cloud Filter | RDAWG 3D/PointCloud | Filter and process point clouds | | Mesh Subdivision | RDAWG 3D/Mesh | Subdivide and smooth meshes | | Neural Renderer | RDAWG 3D/Neural | AI-powered 3D rendering | | Format Converter | RDAWG 3D/Utils | Convert between 3D formats |

⚙️ Configuration

Device Selection

• Auto - Automatically select best device (recommended)
• CUDA - Force GPU acceleration
• CPU - Fallback CPU processing

Memory Settings

• Normal VRAM - Standard memory usage
• High VRAM - Faster processing with more memory
• Low VRAM - Memory-constrained processing

Performance Tips

1. Use CUDA for large models when possible
2. Enable batch processing for multiple objects
3. Adjust mesh resolution based on your needs
4. Monitor VRAM usage with complex scenes

🎯 Examples

Basic 3D Model Viewer Workflow

Load 3D Model → Transform 3D Mesh → 3D to Image → Save Image

Advanced 3D Processing

Load 3D Model → Point Cloud Filter → Mesh Subdivision → Neural Renderer → Enhanced Image

Batch Processing

Load Multiple Models → Batch Transform → Batch Render → Image Sequence

🔧 Troubleshooting

Common Issues

CUDA Out of Memory

• Reduce model complexity
• Use CPU mode for preprocessing
• Enable memory optimization settings

Loading Failures

• Check file format compatibility
• Verify file paths
• Ensure sufficient disk space

Performance Issues

• Update GPU drivers
• Check CUDA installation
• Monitor system resources

Dependencies

Minimum Requirements:

• CUDA 12.8 compatible GPU
• PyTorch 2.9.0+
• Python 3.8+
• 8GB+ VRAM (for large models)

Recommended Setup (Optimal Configuration):

• Python 3.11 (best Open3D compatibility)
• RTX 30-series or newer GPU
• 16GB+ VRAM
• Latest GPU drivers
• SSD storage

🎯 Why Python 3.11?

• Perfect Open3D 0.19.0 wheel support
• Excellent PyTorch CUDA integration
• Most stable for 3D processing libraries
• Better performance than Python 3.13 for scientific computing

🤝 Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details.

🙏 Credits

• PyTorch 3D - Core 3D processing capabilities
• Trimesh - Mesh processing utilities
• Open3D - Point cloud operations
• ComfyUI Community - Framework and inspiration

📞 Support

• GitHub Issues - Report bugs and request features
• Discord - Community support and discussions
• Documentation - Comprehensive guides and tutorials

---

RDAWG 3D Pack - Pushing the boundaries of 3D processing in ComfyUI! 🚀