joone great problem

See more:

Java

dear All
i have a big prroblem in using JOONE it is:
when i designe ANN by joone,i put the values of Lr,momuntum,epochs,training patterns,number of neurones in input,hidden and output layers.the net give some value of error with epochs.the problem is when i remake an ann with the same values of previous net,the net give a differnt values of error.why????
in other words
the same net didn't give the same error when i remake the same net,why???