Thanks, will look through all that. But what I'm ultimately curious about is, since they were trained separately, but on the same data, could the separate training result in different .par files?