MultiLabel Question on Encoding

Hello, I’m trying to follow lesson 6 in fastbook but am having some issues, I’m using the CelebA dataset which has it’s labels hot-one encoded so I tried to do MultiLabelBlock(encoded=True) but am getting the error object of type ‘NoneType’ has no len(). Does anyone have any tips for debugging? I’ve tried changing the input to get_y between series/array but neither works.

Full Code(storage/Untitled.pynb): https://www.paperspace.com/terko65s4/notebook/prxbayrl2

Hi samuel!

Did you fix this error?

I am now working with a similar dataset, multi-label and with one-hod encoded labels. And I’m not sure how to create the DataBlock.

I would appreciate any help about this. Thank you :slight_smile:

@Saioa check out my Kaggle kernel for the toxic comment challenge, as it had one hot encoded labels:

https://www.kaggle.com/muellerzr/end-to-end-fastai2

2 Likes

Thank you very much Zachary! I think it will be very useful to me :slightly_smiling_face:

1 Like