This is an ASR model made using Kaldi. For this model, the resources from Ravnur BLARK 1.0 were modified to comply with the Kaldi guidelines (e.g. the audio files were cut, and they were changed to 16 KHz mono). You can read more on the GitHub page.
You can use Vosk to test the model, just follow the steps below:
- Download the model (click the Download button above).
- Install python
- In your command prompt, type:
- pip install vosk
- pip install sounddevice
- Download this simple python script.
- Run the script in your command prompt (should work for Windows, Mac and Linux).
If you would like to make your own Kaldi model, you can write us at mtd@setur.fo and we can send you the dataset that thas been prepared for Kaldi.
Release: 01.07.2022
Contact: mtd@setur.fo & egil.hofgaard@gmail.com
