We have started to build our own telephony speech dataset, but don’t have a good amount of data yet. It looks like most datasets out there are 16 kHz, does anyone know of other 8 kHz datasets that are available for free or paid? The only one I found so far is:
Hi Nathan,
Have you ben able to create a telephony acoustic model? If so, would you be willing to share it? I have a need to transcribe telephone conversations with DeepSpeech.
Thx.
The best thing to do at this point is to use your own dataset. Probably the best thing is to cross-train from the existing model on your own data.