For those people, who hear about it for the 1st time, JAX is a application method for superior-performance equipment mastering (HPML) research and numerical computing. It is created on the foundation of Python programming language and a commonly acknowledged elementary package NumPy which is utilised for scientific computing in the Python ecosystem.
JAX supports the hardware acceleration, just-in-time compiling your personal Python capabilities, running NumPy courses on several-core GPU/TUP (i.e. graphical and tensor processing models). Thanks to a sophisticated framework it gives its buyers with the possibility to outline and manipulate custom practical transformations, expressing elaborate algorithms and