PhiML

🌐 GitHub     🔗 API     ▶ Videos   •   Introduction   •   Examples

We recommend starting with the introduction notebook which walks you through the installation and introduces the general structure of Φ_ML. Check out the corresponding introductory ▶ tutorial video as well!

Tensors & Math

Getting started

• Shapes
• Φ_ML’s Tensor
• Data types

If you’re already familiar with NumPy, PyTorch, TensorFlow, or JAX

• Advantages of Dimension Names and Types
• Advantages of Precision Management

Advanced topics

• Writing n-dimensional code
• Matrices and automatic matrix generation
• Linear solves
• Non-uniform tensors
• NumPy for constants
• Selecting compute devices (CPU/GPU/TPU)
• Automatic differentiation
• Just-in-time (JIT) compilation
• Converting between Jax, PyTorch, TensorFlow, NumPy
• What to avoid in Φ_ML
• Performance

Neural Networks

• Training neural networks
• Neural Network API

API Documentation

The 🔗 API documentation is generated using pdoc3. To manually generate the documentation, add the Φ_ML directory to your Python path and run

$ pdoc --html --output-dir docs --force phiml

This requires PyTorch, TensorFlow and Jax to be installed, in addition to the standard Φ_ML requirements.

Contributing to Φ_ML

Contributions are welcome!

If you find a bug, feel free to open a GitHub issue or get in touch with the developers. If you have changes to be merged, check out our style guide before opening a pull request.

This site is open source. Improve this page.