DeepSpeech for Jetson/Xavier

In the Mozilla forum there is some older discussion that suggests that the DeepSpeech architecture does not support phoneme transcription:

Last year there was another discussion thread where a community member did some experiments with phoneme transcription with - to my understanding - mixed results:


I have build a standalone version of the DeepSpeech ctc-Decoder, maybe this is of helpful for someone:

I did an evaluation of freely available STT models for german language: