I noticed that in your code the loss function needs labels. In your paper you mentioned that "GraphRNA could be trained with an unsupervised, supervised, or semi-supervised setting". In my case, the data labels are N/A, any suggestion on how to define the loss function ? Thanks.
I noticed that in your code the loss function needs labels. In your paper you mentioned that "GraphRNA could be trained with an unsupervised, supervised, or semi-supervised setting". In my case, the data labels are N/A, any suggestion on how to define the loss function ? Thanks.