[Kaldi-users] training recommendataions

[Mailing l. u. f. U. C. a. U. <kal...@li...>]

I was curious likely logical next-steps for improving training / decoding might be give other user's experience.  It does very well with significant background noise, but I think that there is definitely room to do better.  

1 - I have a training set of around 5K words, though I could bring that up to around 20K
2 - I am using the kaldi_lm, though I could use SRILM . . not sure if it would necessarily improve results
3 - I am decoding about 1 minute of text, though training data is in 10 second epochs.  I can mix some of the test data in if that would help. 
4 - When I am training deltas I use a very small # of leaves / gauss (100 / 1000) to get the best results.   The best results are with tri1.  Further training yields worse results.  
5 - I use the same lexicon for the training and decoding (though a more restrictive language model for decoding).  

Any help / thoughts are appreciated.  

Thanks,

Nathan