search

srvk / eesen

The official repository of the Eesen project
http://arxiv.org/abs/1507.08240
Apache License 2.0
824 stars 342 forks source link

Clarification Regarding Using WFST decoding #210

Closed rmalav15 closed 5 years ago

rmalav15 commented 5 years ago

Hey,

Much thanks for this very useful project. After following this lecture slide, I was able to generate final decoding graph TLG.fst. For my project requirement, I want to use the WFST decoding with CTC (char label) generated by external software such as DeepSpeech. The scripts given in examples for such task also uses the Essen Acoustic Model in loop (to the best of my knowledge), which I dont require. Can you please help me to look into the right direction for writing a script for my task?

Thanks, Ram

rmalav15 commented 5 years ago

Found it. https://github.com/srvk/eesen/issues/148. Thanks.