Training offline speech-recognition software with half-transcribed interviews

72 Views Asked by At

I searched a bit for a speech recognition software that would work offline and allow me to train it with half-finished transcripts. It seems that Mozilla's deepspeech could do the job. The issue of the interviews that I have to transcribe is that there are two voices and non-standard English accents. Does anyone have any experience with DeepSpeech in order to do this and could guide me to a tutorial? Or would you recommend another software (preferently offline and opensource) ?

0

There are 0 best solutions below