Matlab Coding in C++ with forest is very long and consumes it in RAM

We try to generate c++ library from Matlab functions with random forest inside.

We run code generation on 2,4ghz cpu and 60 Go RAM.

Our 3 forests size (50-120Mo).

The library generation duration is 3 hours and it use 60go RAM.

The other library code generation duration are a few minutes (without random forest of course).

My questions :

Do you have similar codegen duration with random forest? Same question with RAM usage?

Do you have tricks to speed up code generation and decrease RAM usage?

Answer 1

[Manish]

Hi,

Regarding the query, "Do you have tricks to speed up code generation and decrease RAM usage?":

You can use the 'compact' function to reduce the size of the model.

Refer to the documentation below for more information:

• https://www.mathworks.com/help/stats/regressionchainensemble.compact.html

Additionally, please refer to the following documentation for steps to increase code generation speed:

• https://www.mathworks.com/help/ecoder/ug/increase-code-generation-speed.html

Hope this helps!