Hi, I followed the doc exactly to test what the problem is, but the dataset loaded still recognize audio as a feature only, the metadata and struct look like this:
Do you know how to fix this problem? Thanks
