MLJ.jl

"Is it good?" Compared to the CUDA dumpster fire at a cat food factory, it is often trivial and nearly transparent syntax for the users familiar with ML. Really depends on the use-case =) The conventional options are fairly well documented: https://sciml.ai/ https://fluxml.ai/Flux.jl/stable/gpu/ https://github.com/SciML/DiffEqFlux.jl https://github.com/alan-turing-institute/MLJ.jl

Then you haven't looked at Julia's ecosystem. It may not be quite as mature, but it's getting there quickly. It's also far more interoperable because of Julia's multiple dispatch and abstract types. For example, the https://github.com/alan-turing-institute/MLJ.jl ML framework (sklearn on steroids), works with any table object that implements the Tables.jl interface out of the box, not just with dataframes. That's just one example.

It's seems like a new competing approach in the ML space. The standard julia data pipeline right now usually involves either directly using an independent package for each step (or rolling your own since Julia has strong native support for fast data/number manipulation) or using a package that combines all those functionalities like MLJ.jl [1] for ML and Queryverse.jl [2] for data manipulation and visualization. The advantage is that since Julia lang focuses on composability, those frameworks can easily share core functionality so library creators don't need to reinvent the wheel and all frameworks can benefit from each other works. [1] https://github.com/alan-turing-institute/MLJ.jl [2] https://www.queryverse.org/