The Whisper models 100h are acoustic models suitable for Automatic Speech Recognition in Faroese. They are the result of fine-tuning different “openai/whisper” models with 100 hours of Faroese data released by the Ravnur Project from the Faroe Islands and curated by Carlos Mena during March (2023) in the servers of the Language and Voice Lab at Reykjavík University in Iceland. The specific dataset used to create the model is this one.
Below you can see the different versions of the Whisper accoustic models trained on different versions of the openai Whisper models.
Large v2 8k Large 8k Large 8k ct2
Release: 2022
Contact: carlos.mena@ciempiess.org
