> Machine learning requires alot of experimentation and C++ is a terrible medium for that.
Is it really? I suspect you do not know how to use C++ effectively. Why not try parameterizing your machine learning algorithms with templates? That way, you can plug in different loss functions, different regularizers, different metrics, or different initialization routines without having to modify your internal algorithm code. In addition, because you are using templates, it is bound at compile time, so you don't pay the vtable lookup penalty that you do with inheritance.
Python might be better for your uses, and I can respect that, but don't tell people that C++ is a terrible way to implement machine learning algorithms when you actually are just ignorant of how to do it well.
> Machine learning requires alot of experimentation and C++ is a terrible medium for that.
Is it really? I suspect you do not know how to use C++ effectively. Why not try parameterizing your machine learning algorithms with templates? That way, you can plug in different loss functions, different regularizers, different metrics, or different initialization routines without having to modify your internal algorithm code. In addition, because you are using templates, it is bound at compile time, so you don't pay the vtable lookup penalty that you do with inheritance.
Python might be better for your uses, and I can respect that, but don't tell people that C++ is a terrible way to implement machine learning algorithms when you actually are just ignorant of how to do it well.