v1.0.3.0

Multithreading is here!
Added multithreading support and related with this functions.
Added function to check if the processor supports AVX instructions. (Available in class "NeuralUtils")
Added function to add randomly noise into float arrays. (Available in class "NeuralUtils")
Added function to fill randomly float arrays. (Available in class "NeuralUtils")
Added utility function to copy from one float array to another. (Available in class "NeuralUtils")
Updated function of saving neural network to file. (Layers configurations now fully saves.)
Others small changes.

v1.0.2.2

Added function to calculate the mean square error and root mean square error.
Renamed function overload "ForwardPropagate" that using pointers (and not creating array for output) to "ForwardPropagatePtr".

v1.0.2.1

Fixed incorrect image conversion when the image width or height is not even 4.

v1.0.2.0

Added new experimental image convertion algorithm "YUVI". Based on YUV color model, but UV (color diff) indexed into 1 channel. (Brightness channel + Indexed UV (color) channel)
Fixed incorrect loading neural network from file. (Due to typos in code)
Fixed working of function "GetSubNeuralNetwork".
Added feature to forward propagate input data to first layer output. (Specially for auto-encoders with separated encoder and decoder)

v1.0.1.0

Changed library namespace from "Rio_Neural_Network" to "RioNeuralNetwork".
Changed method of converting RGB image to grayscale. (Luminosity algorithm)
Renamed Utils class to NeuralUtils. (frequently used name, so as not to conflict with "using RioNeuralNetwork;")
Added support of converting 24bit per pixel images.
Added function to make sub neural network from full. (Acts like as substring but for neural network layers. Very helpful for getting encoder and decoder from not separated autoencoder)

v1.0.0.0

First release