Unfolding of recurrent neural nets for back-propagation training

When training a layrecnet (I believe this is the right function to use for a "standard" recurrent neural network), and training e.g. by the default trainlm method, how many time-steps are unfolded for the back-propagation? Can this number be specified and if so, how?

Answer 1

[Greg Heath]

You question doesn't make any sense to me. Have you

1. Read the documentation

 help layrecnet

 doc layrecnet

2. Searched "layrecnet" in previous postings of both the NEWSGROUP and ANSWERS?

Hope this helps.

Greg