I work with kaldi just some months, and have a qst about the
WER_decoding result.
I test my own data, mainly by the run.sh script, and i get puzzling result:
i get WER relatively high in the train_decoding (about 20%), and very
low WER in the tset_decoding (about 3%). Can someone help me?
Beacuse i am new user, please explain the answer clearly.
Thank you very very much
A. tuk
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi all
I work with kaldi just some months, and have a qst about the
WER_decoding result.
I test my own data, mainly by the run.sh script, and i get puzzling result:
i get WER relatively high in the train_decoding (about 20%), and very
low WER in the tset_decoding (about 3%). Can someone help me?
Beacuse i am new user, please explain the answer clearly.
Thank you very very much
A. tuk