✨ Features: - Modern 3D processing nodes for ComfyUI - CUDA 12.8 + PyTorch 2.9.0 optimization - Open3D 0.19.0 integration with Trimesh fallback - GPU-accelerated mesh processing and rendering - Complete installation and development tooling 📦 Package Contents: - 4 core 3D processing nodes - Comprehensive documentation - Example workflows (Basic Viewer, Product Photography) - Development scripts and tools - Automated dependency management 🚀 Ready for production use with ComfyUI! 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
5.0 KiB
5.0 KiB
Installation Guide
Quick Start
Method 1: ComfyUI Manager (Recommended)
- Open ComfyUI Manager
- Search for
rdawg-3d-pack-cu128-pytorch-2.9.0 - Click Install
- Restart ComfyUI
Method 2: Manual Installation
cd ComfyUI/custom_nodes
git clone https://github.com/rdawgemfl/rdawg-3d-pack.git
cd rdawg-3d-pack
python install.py
Method 3: Development Installation
git clone https://github.com/rdawgemfl/rdawg-3d-pack.git
cd rdawg-3d-pack
python scripts/setup_development.py
System Requirements
Minimum Requirements
- Python: 3.8 or higher
- PyTorch: 2.9.0 or higher
- CUDA: 12.0 or higher (for GPU acceleration)
- VRAM: 4GB+ (8GB+ recommended for large models)
- RAM: 8GB+ (16GB+ recommended)
Recommended Setup
- Python: 3.11 (best compatibility)
- PyTorch: 2.9.0+cu128
- GPU: RTX 30-series or newer
- VRAM: 16GB+
- SSD Storage
Dependency Installation
Core Dependencies (Required)
pip install torch>=2.9.0 torchvision>=0.24.0
pip install numpy>=1.24.0
pip install trimesh>=4.0.0
pip install tqdm>=4.65.0
pip install einops>=0.7.0
Optional Dependencies
Full 3D Processing Suite
pip install open3d>=0.18.0
pip install pytorch3d>=0.7.5
pip install scipy>=1.11.0
pip install matplotlib>=3.7.0
Neural Rendering Capabilities
pip install transformers>=4.35.0
pip install diffusers>=0.24.0
pip install accelerate>=0.24.0
pip install xformers>=0.0.32
Platform-Specific Installation
Windows
Using Python Embedded (Recommended for ComfyUI)
# Navigate to your ComfyUI Python installation
cd "D:\ComfyUi\Cuda-v12\python_embeded"
# Install core dependencies
python.exe -m pip install torch>=2.9.0 torchvision>=0.24.0 --index-url https://download.pytorch.org/whl/cu128
python.exe -m pip install trimesh>=4.0.0 open3d>=0.18.0
Using System Python
# Install PyTorch with CUDA support
pip install torch>=2.9.0 torchvision>=0.24.0 torchaudio>=2.9.0 --index-url https://download.pytorch.org/whl/cu128
# Install 3D libraries
pip install trimesh open3d
Linux
# Install PyTorch
pip install torch>=2.9.0 torchvision>=0.24.0 torchaudio>=2.9.0 --index-url https://download.pytorch.org/whl/cu128
# Install system dependencies for 3D processing
sudo apt-get update
sudo apt-get install libegl1-mesa-dev libgl1-mesa-dev
# Install Python packages
pip install trimesh open3d pytorch3d
Troubleshooting
Common Issues
PyTorch CUDA Issues
# Check CUDA availability
python -c "import torch; print(f'CUDA available: {torch.cuda.is_available()}')"
# Reinstall PyTorch if needed
pip uninstall torch torchvision torchaudio
pip install torch>=2.9.0 torchvision>=0.24.0 --index-url https://download.pytorch.org/whl/cu128
Open3D Installation Issues
# For Windows: use pre-built wheel
python -m pip install https://github.com/isl-org/Open3D/releases/download/v0.19.0/open3d-0.19.0-cp311-cp311-win_amd64.whl
# For Linux: build from source if wheel fails
pip install open3d --verbose
Memory Issues
- Reduce model complexity for large meshes
- Use CPU mode if VRAM is limited
- Enable memory optimization in settings
Import Errors
# Check Python path
python -c "import sys; print(sys.path)"
# Reinstall package
pip uninstall rdawg-3d-pack-cu128-pytorch-2.9.0
pip install -e .
Performance Optimization
GPU Memory Management
# In your ComfyUI settings or environment variables
import os
os.environ['PYTORCH_ALLOC_CONF'] = 'max_split_size_mb:128'
Multi-GPU Setup
# Specify GPU device
import torch
torch.cuda.set_device(0) # Use first GPU
Verification
Test Installation
# Run the built-in test suite
python scripts/test.py
# Or test manually
python -c "
import torch
import trimesh
try:
import open3d as o3d
print('✅ Open3D available')
except ImportError:
print('⚠️ Open3D not available')
print('✅ Installation successful!')
"
Test in ComfyUI
- Start ComfyUI
- Look for "RDAWG 3D" nodes in the node menu
- Try loading a simple 3D model
- Check console for loading messages
Configuration
Environment Variables
# Optional: Set default device
export RDAWG_3D_DEVICE=cuda # or cpu
# Optional: Memory optimization
export PYTORCH_ALLOC_CONF=max_split_size_mb:256
ComfyUI Settings
Add to your ComfyUI configuration:
{
"rdawg_3d_pack": {
"default_device": "auto",
"memory_optimization": true,
"open3d_priority": true
}
}
Next Steps
After successful installation:
- Browse Examples: Check the
examples/directory for workflow examples - Read Documentation: See
docs/USAGE.mdfor detailed usage instructions - Join Community: Get support and share your creations
- Contribute: Help improve the package by reporting issues or submitting pull requests
Support
- GitHub Issues: Report bugs and request features
- Documentation: Check
docs/for detailed guides - Community: Join discussions and get help