This model performs verbatim ASR. It is trained to jointly perform verbatim ASR and automatic subtitling, after which the subtitle part has been discarded. It is trained on Flemish Dutch speech, and can be used with the standard ESPnet toolkit.

Version: September 2023