News for hackers. Stuff that matters.
Re:MLJ.jl: A Julia package for composable machine learning by Skyy93 on Monday April 12, 2021 @04:40AM (#26777168)
I do not understand why someone should use this? It is basically a wrapper of existing Libs and Frameworks. This framework does not solve any problem that scikitlearn and other existing frameworks have not already solved.
Re:MLJ.jl: A Julia package for composable machine learning by sgt101 on Monday April 12, 2021 @05:42AM (#26777564)
A good thing to read is https://joss.theoj.org/papers/10.21105/joss.02704

The focus of this package is on the "plubming" of auto-ml solutions. The work of designing solutions based on multiple discovered models is rather unsupported in the current state of the art, and this package looks quite supportive of it.

A big contribution (if it works in practice) is the idea of a scientific type describing what data is and how it should be mapped to algorithms - a systematic way of doing this is the underpinning of a process for model selection.

Re:MLJ.jl: A Julia package for composable machine learning by dklend122 on Monday April 12, 2021 @06:33AM (#26777887)
Check out the " Model composability" section. Also multithreading, custom differentiable loss functions, works with any abstract table type, and the list goes on