Multi Label Classification in pytorch

[AjayTalati]

Hi there,

I can’t get multi-label classification working either, but @bartolsthoorn and @mratsim have found possible ways to do it here

Calculating accuracy for a multi-label classification problem

I came up with the answer: torch.sigmoid(classified_labels).data > 0.5 will give the correct labels with MultiLabelSoftMarginLoss().

hope that helps?

Aj