Spaces:

m-a-p

/

Music-Descriptor

like 20

Running

I notice that the output feature from the model will be extracted from first layer (https://huggingface.co/spaces/m-a-p/Music-Descriptor/blob/3462d760643cbc4b9c7a11f5eaec4915aa468470/app.py#L152),
but it is different from the extracting way in Music-Genre-Tagging (https://huggingface.co/spaces/m-a-p/MERT-Music-Genre-Tagging-Prediction/blob/82aecc2e4783f68a23fa68f067784ae0659c73a1/app.py#L123)

I want to make sure that will not affect the indexing for downstream task.
Thanks!