Installation
Requirements
BSBR requires Python 3.12 or higher and PyTorch 2.6.0 or higher. The package is designed to work with modern deep learning frameworks and tools.
Basic Installation
The simplest way to install BSBR is using pip:
Development Installation
For development or to use the latest features, you can install directly from GitHub:
# Clone the repository
git clone https://github.com/JacobFV/bsbr.git
cd bsbr
# Install in editable mode
pip install -e .
Optional Dependencies
BSBR provides several optional dependency groups that you can install:
# Install with all extras (evaluation tools, visualization, etc.)
pip install "bsbr[extras]"
# Install with documentation tools
pip install "bsbr[docs]"
# Install with all optional dependencies
pip install "bsbr[all]"
Available Extras
extras
: Evaluation tools, visualization utilities, and research componentsdocs
: Documentation building tools and dependenciesall
: All optional dependencies
GPU Support
BSBR automatically uses GPU acceleration when available. To ensure GPU support:
-
Install CUDA-enabled PyTorch:
-
Verify GPU availability:
Troubleshooting
Common Issues
- Import Errors
- Ensure you're using Python 3.12 or higher
- Check that PyTorch is installed correctly
-
Verify your Python environment is activated
-
CUDA Errors
- Confirm CUDA toolkit is installed
- Verify PyTorch CUDA version matches your system
-
Check GPU drivers are up to date
-
Memory Issues
- Adjust batch size or sequence length
- Enable gradient checkpointing
- Use smaller model configurations
Getting Help
If you encounter issues:
- Check the GitHub Issues
- Join our Discord Community
- Create a new issue with:
- Python version
- PyTorch version
- Error message
- System information
Next Steps
After installation, you can:
- Try the Quick Start Guide
- Explore the Examples
- Read the User Guide