For large, fast, and easy numeric computation, JAX is an alternative to Pytorch and Tensorflow. JAX has all the benefits such as acceleration and scalability on, and across, devices without the boiler plate code. This allows you to focus on the maths without sacrificing speed. This talk provides a gentle introduction to JAX. We will also go over several other packages in the JAX ecosystem. These packages, such as Haiku, allow you to focus on the high-level concepts or help you catch nasty bugs.
About Hylke Donker
Data scientist at University Medical Centre Groningen.