3 ms·
> at each layer of the neural work This article is poorly written, with mistakes and lack of domain knowledge. > in deep learning, when functions are applied
by visarga 7y ago
> at each layer of the neural work
This article is poorly written, with mistakes and lack of domain knowledge.
> in deep learning, when functions are applied to the data, the data does not change
Except for in-place operations.
> Clojure doesn’t replace the Java thread system, rather it works with it. Since the core data structures are immutable, they can be shared readily between threads
Missing the point - basic ops in DL need to be optimised to an extreme level, and that implies using C.
- fock 7y agoso I'm not the only one wondering how some threading examples (in Clojure?) generalize to (suggested) applicability of these language for ML...