Machine Learning WG updates - Sep 2021

Welcome to another update from the Machine Learning WG! This time we are glad to announce many improvements to our on-going efforts and welcome new projects into the fold!

Nx updates

Nx is a multi-dimensional tensors library for Elixir with multi-staged compilation to the CPU/GPU.

  • Refactored Nx’ core to support TPUs (

  • Many new index-based operations were added, such as Nx.take/3, Nx.take_along_axis/3, Nx.gather/2, and Nx.scatter_add/3

  • The EXLA backend (which uses Google’s XLA) has been updated to Tensorflow 2.6 and now supports NVIDIA’s Jeton and macOS aarch64

  • We now precompile Google’s XLA to multiple OSes and architectures. This brings the getting started experience down from several minutes to a couple seconds

Explorer updates

We are glad to welcome Explorer, led by Chris Grainger, into the Machine Learning WG and into the elixir-nx organization. Explorer is a dataframe library for Elixir. It is an API for data manipulation with high-level features such as:

  • Simply typed series: float, integer, boolean, string, date, and datetime

  • A powerful but constrained and opinionated API, so you spend less time looking for the right function and more time doing data manipulation

  • Pluggable backends, providing a uniform API whether you’re working in-memory or (forthcoming) on remote databases or even Spark dataframes

  • The first (and default) backend is based on NIF bindings to the blazing-fast polars library

Learn more about it in its README or in its introductory notebook.

Livebook updates

Livebook is a collaborative and interactive code notebook in Elixir developed by Jonatan Kłosko.

Livebook has been moved to its own organization and it will launch their own website soon. The organization also hosts Nerves Livebook, led by Frank Hunleth, which provides a straightforward way to run Livebook on embedded devices and ships with many additions to teach Nerves to new developers through Livebook.

Axon updates

Axon is a library led by Sean Moriarity that brings deep learning to Elixir. It is powered by Nx and runs on the CPU/GPU.

Other bits

Are you also interested in Machine Learning and the Erlang Ecosystem? Join us and come chat on our Slack.