In the new Audio Course, under the ‘Fine-tuning a model for music classification’, the code in the first block should be updated from (loading it as DatasetDict):
gtzan = load_dataset(“marsyas/gtzan”, “all”)
to (load it as Dataset):
gtzan = load_dataset(“marsyas/gtzan”, split=“all”)
Otherwise the second block of code would throw as error:
AttributeError: ‘DatasetDict’ object has no attribute ‘train_test_split’