A neural network library for Elixir built on Nx, providing functional, model creation, and training APIs for deep learning.
Axon is a neural network library for Elixir that provides tools for building, training, and deploying deep learning models. It solves the problem of bringing modern deep learning capabilities to the Elixir ecosystem by offering functional, model creation, and training APIs powered by Nx for efficient numerical computation.
Elixir developers and data scientists who want to implement deep learning models within the Elixir ecosystem, leveraging its concurrency and fault-tolerance features for machine learning workloads.
Developers choose Axon because it integrates seamlessly with Elixir and Nx, offering a clean API inspired by PyTorch, full control over model components, and the ability to compile models using various Nx backends for performance optimization.
Nx-powered Neural Networks
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Seamlessly leverages Elixir's concurrency and fault tolerance for ML workloads, making it ideal for integrating deep learning into existing Elixir applications, as emphasized in its philosophy.
Built entirely on Nx's defn, enabling JIT/AOT compilation with backends like EXLA for GPU acceleration, as highlighted in the Functional API section.
Training API models PyTorch Ignite, offering familiar event handlers and metrics for streamlined model training, detailed in the Training API description.
Models are Elixir structs with straightforward serialization and visualization via Axon.Display, providing clear summaries and future format compatibility, as shown in the Model Creation examples.
Admits it aims to match other frameworks but may lack advanced layers or features, relying on community contributions, as noted in the Model Creation section where it encourages opening issues for missing functionality.
Practical use requires additional compilers like EXLA, adding setup complexity and potential compatibility hurdles, as specified in the Installation instructions.
As a newer project in the Elixir ML space, it has fewer tutorials, pre-trained models, and third-party tools compared to established Python libraries, impacting ease of adoption.
Axon is an open-source alternative to the following products:
TensorFlow Keras is the high-level neural networks API integrated into TensorFlow, providing an easy-to-use interface for building and training deep learning models.
PyTorch is an open-source machine learning framework that provides tensor computation with strong GPU acceleration and deep neural networks built on a tape-based autograd system.