HELP using trained models with decoder

[Alexssandra Stefani]

Hi,

 i was wondering if i would be able to get a bit of guidance with the trained NN models. i have successfully trained a small corpus using the 'NNScripts' provided but am a little lost in knowing how to use them with the phnrec. 

When the training finishes  i get a folder of the last iteration of training which contains other folders such as weights,norms.dict etc etc.  from a previously answered question (https://groups.google.com/forum/#!topic/phnrec/W7QrSN5MFtY) i understandthat  we should rename the weights and norms files as streams_s0_1-> band0.weights?

 However i have no .weights or .norms files in my weights and norms folders  respectively, but files such as merger_norm.ftr ,stream1_norm.ftr,stream2_norm.ftr in my norm folder and full_sys.ftr,full_sys_fwd.ftr,merger.ftr,stream1_fwd.ftr,stream2_fwd.ftr,streams.ftr,streams_fwd.ftr in my weights folder. These are also a different format as the .weights and .norms files that come in the pre trained models such as the TIMIT.

i did attempt at renaming a few of these and using it with the phnrec decoder but it was throwing an error in concern to the .weights and .norm files.

Any help would be greatly appreciated

Thanks

Alessandro