I have got a simple [Nemotron ASR](https://build.nvidia.com/nvidia/nemotron-asr-streaming/deploy) pipeline working (remote linux server with an RTX5090, Windows client) and it's pretty fast and quite accurate. It would be good to incorporate as a Dictation option for the Kaldi backend.
I have got a simple Nemotron ASR pipeline working (remote linux server with an RTX5090, Windows client) and it's pretty fast and quite accurate.
It would be good to incorporate as a Dictation option for the Kaldi backend.