Everardo Robredo
2009-09-30 22:15:17 UTC
Hi,
I've been working with cascade training and for all the training patterns
that I build, the algorithm always constructs a network with the following
structure:
k,1,1,1,...,1,1,1,r
where 'k' and 'r' are the sizes of the input and output layers, which are
taken from the data file I provide. This means that all the neurons are
added in different layers giving a chain-like topology so if N neurons are
added, I end up with N hidden layers.
Since I'm new with FANN, and specially with cascade algorithm, I wonder if
this has to be with my code, which by the way is almost the same as the
cascade example code.
I've been working with cascade training and for all the training patterns
that I build, the algorithm always constructs a network with the following
structure:
k,1,1,1,...,1,1,1,r
where 'k' and 'r' are the sizes of the input and output layers, which are
taken from the data file I provide. This means that all the neurons are
added in different layers giving a chain-like topology so if N neurons are
added, I end up with N hidden layers.
Since I'm new with FANN, and specially with cascade algorithm, I wonder if
this has to be with my code, which by the way is almost the same as the
cascade example code.
--
Everardo
Everardo