Using Language Models to decode.

gooofy / zamia-speech

Open tools and data for cloudless automatic speech recognition

GNU Lesser General Public License v3.0

443 stars 84 forks source link

Using Language Models to decode. #119

Open IanMGriff opened 2 years ago

IanMGriff commented 2 years ago

Hello,

Thank you for sharing all your fantastic work! I was wondering if it would be possible to use an existing language model to decode the outputs of a pytorch model that is trained to predict phonemes (CMU pronunciation dictionary phones) from speech.

I am fairly new to ASR, so I am sorry if this is obvious or something already solved.

Thanks!