Are participants allowed to pretrain or fine-tune models on adult speech data first, and then adapt those models to child speech for this competition?
Yes, you are allowed to train models with external data as long as doing so meets the external data requirements. Please refer to the home page and rules for more details.
Hey @hannahmoro! Can you provide a specific list of dialects/accents/backgrounds that might broadly cover most of your test-set as described in the problem statement?
The challenge dataset encompasses a broad range of U.S. regional dialects, accents, and socio-linguistic backgrounds …
Hello! Unfortunately, we cannot provide specific demographic details of the data in the test set. Please continue to refer to the problem description for details on the dataset.