Yes, you need to set push_to_hub=True in the training arguments before training. Check the updated notebooks!
push_to_hub=True