Multi-class classification with Multi-hot encoded vector

Beginners
1

Hi, I am very new to the huggingface community and a newbie. :smiley:

In my Sentiment Analysis training set I have a multi-hot encoded vector for the labels, where each 1 represents the existance of the label: [1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0].

I am trying to proceed with some tutorials based provided by Huggingface but to my knowledge there seems to be nothing regarding my problem.

# train_ds:
Dataset({
    features: ['work_id', 'labels', 'text'],
    num_rows: 307102
})

# train_ds.features: 
{'work_id': Value(dtype='string', id=None), 

'labels': Sequence(feature=ClassLabel(names=['pornographic-content', 'violence', 'death', 'sexual-assault', 'abuse', 'blood', 'suicide', 'pregnancy', 'child-abuse', 'incest', 'underage', 'homophobia', 'self-harm', 'dying', 'kidnapping', 'mental-illness', 'dissection', 'eating-disorders', 'abduction', 'body-hatred', 'childbirth', 'racism', 'sexism', 'miscarriages', 'transphobia', 'abortion', 'fat-phobia', 'animal-death', 'ableism', 'classism', 'misogyny', 'animal-cruelty'], id=None), length=-1, id=None), 

'text': Value(dtype='string', id=None)}

Could someone help me, could I am not really sure how to proceed

Best thanks,
Rebecka

2

Have a similar issue, did you figure it out?

3

There are options for binary multi label problems but not for multi class tasks with multiple hot encoded labels