Installation

PyTorch Graph can be installed using pip. The package is available on PyPI and supports Python 3.8+.

Basic Installation

Install the core package:

pip install pytorch-graph

This will install PyTorch Graph with all essential dependencies.

Enhanced Installation

For additional features and better performance:

pip install pytorch-graph[full]

This includes: * Enhanced visualization features * Additional export formats * Performance optimizations * Extended color schemes

Development Installation

For development and contributing:

pip install pytorch-graph[dev]

This includes: * Development dependencies * Testing frameworks * Code quality tools * Documentation tools

From Source

Clone the repository and install in development mode:

git clone https://github.com/your-username/pytorch-graph.git
cd pytorch-graph
pip install -e .[dev]

Requirements

Core Requirements

Python: ≥ 3.8
PyTorch: ≥ 1.8.0
matplotlib: ≥ 3.3.0
numpy: ≥ 1.19.0

Optional Requirements

plotly: For interactive visualizations
torchinfo: For enhanced model summaries
networkx: For advanced graph analysis
pillow: For image processing

Verification

Verify your installation:

import torch
from pytorch-graph import generate_architecture_diagram

# Create a simple model
model = torch.nn.Sequential(
    torch.nn.Linear(10, 5),
    torch.nn.ReLU(),
    torch.nn.Linear(5, 1)
)

# Generate a test diagram
generate_architecture_diagram(
    model=model,
    input_shape=(1, 10),
    output_path="test_diagram.png"
)

print("✅ PyTorch Graph installed successfully!")

Troubleshooting

Common Issues

ImportError: No module named ‘torch’

Install PyTorch first: pip install torch

ImportError: No module named ‘matplotlib’

Install matplotlib: pip install matplotlib

Permission denied errors

Use pip install --user pytorch-graph for user installation

Version conflicts

Use a virtual environment:

python -m venv pytorch-graph-env
source pytorch-graph-env/bin/activate  # On Windows: pytorch-graph-env\Scripts\activate
pip install pytorch-graph

Performance Tips

Use pytorch-graph[full] for better performance
Ensure you have sufficient memory for large models
Use GPU acceleration when available
Consider using smaller input tensors for initial testing

Support

If you encounter issues:

Check the troubleshooting section
Search GitHub Issues
Create a new issue with detailed information
Join our Discussions