Are there alternatives to the FANN library that use computer resources to the fullest?

Question

ruby

The situation is as follows: I wrote in ruby ​​using ruby_fann, the PC has an overclocked Core i7-6700K on board, but the library uses only one processor core during training, as a result, neuron training takes a very long time.
Is it possible to force the library to use the available resources?
Perhaps it is worth taking some other implementation where possible?
Do such things lack the ability to parallelize the process?

Answer 1

FANN is almost never used in python - they usually use keras, although it is slightly slow compared to MXnet.
If speed is important, then use MXnet.
MXnet can use all the resources of the computer, including the video card.