Clojurians Log v2

> I’m learning ML Depends on how much performance you need. For non neural networks (e.g. decision trees) - these are mostly implemented in CPU - you may even be able to just using normal clojure types (vectors, maps etc). For small nerual neworks (e.g. less than a 100,000 nodes) you can do these again using base clojure types or something more optimised like dtpe-next. For bigger neural networks you will likely want GPU support.

check "noj" for ML (non deep-learning) https://scicloj.github.io/noj/

Use neanderthal - then moving to the GPU is easy.

Or use a preexisting python library via libpython-clj. dtype-next in this situation is the interop language and provides some processing capabilities but not really efficient dense numeric operations.