Is it possible to get the ground truth transcripts for the dataset, so one doesn't have to use ASR for multimodal fusion training. I am not able to locate any.
· Sign up or log in to comment
