Weights to hidden neurons in neural network -


How are the weight given between input-neurons and hidden-neurons and hidden-neurons and output -neurons ? I know that in the beginning weighs randomly.

Secondly, I am recognizing the character and say that I have a size of 64 input neurons which is a prototype of 8x8 pixels, it should mean that I should have 64 output right with neurons. the wanted?

See my answer for the size of the output layer is the only question.

I am sure how "weight is given" means what it is. Do you mean "trained"? If so, usually by backpropegation If you "represent it" means: usually in the form of an array or matrix.

If you want to read more about fine-tuning for backspace, then read by LeCun.

On another note: The pixels of 1 node are in the form of input that you do never You never call raw data in the network because it contains noise and unnecessary information is. Find a representation, a model, an encoding or something similar before you feed it in the network. To understand how this is done, you have no other choice than doing some research. There are many possibilities to give a clear answer.


Comments

Popular posts from this blog

ios - How do I use CFArrayRef in Swift? -

eclipse plugin - Run java code error: Workspace is closed -

c - Error on building source code in VC 6 -