ml-for-speech/edacc

Reformatted version of https://huggingface.co/datasets/edinburghcstr/edacc for audio classification

Make sure to remove the "Don't know" label:

dataset = dataset.filter(lambda example: example["label"] != 15)