david-estevez / neuralnetwork Goto Github PK

The code is now undocumented, but documentation is a requirement for this proyect.

Load matrix function, belonging to NNFileInput, currently loads matrix data row by row.

This could be done faster by loading it and storing the data element by element on the matrix.

All matrices that are passed to other functions are copies of all the data of the layer/network. If you are working with relatively big networks, that may result inefficient from the memory management point of view.

Matrices with pointers/references should be implemented.

No layer takes into account the bias unit. They should.