Dataset for multilabel classification

[celalp]

Hi I have a dataset for multi-label classification, I have 14 categories and a decent amount of data. However I’m having issues trying to get the training going.

I have tried one hot encoding or passing the list of classes but keep getting errors.

Basically I have a dataset that looks like this:

{"text":["text1", "text2", "text3"], "labels":[[0,0,0], [1,1,1], [0,1,3]]}

I have been searching a lot of places and there are conflicting things and virtually no documentation on this subject. Does anyone have any code snippets for creating datasets or training such models?

Thanks.

[John6666]

I hope this is helpful…

Dataset label format for multi-label text classification

Hi, I’ve been able to train a multi-label Bert classifier using a custom Dataset object and the Trainer API from Transformers. The Dataset contains two columns: text and label. After tokenizing, I have all the needed columns for training. For multi-label classification I also set model.config.problem_type = "multi_label_classification", and define each label as a multi-hot vector (a list of 0/1 values, each corresponding to a different class). So for instance, given the task consists in assig…

Multi-Label Classification Model From Scratch: Step-by-Step Tutorial