Is there a way to train a multi-label classification model on multi-class labeled dataset by manipulating this dataset in anyway or by any other possible way? @stanfordnlp @ML_NLP @uwnlp @ucl_nlp

@nerd_nlp @ML_NLP @uwnlp @ucl_nlp Sure. The simplest and commonest way is to have one binary target bit for each class, as in this Keras demo, though there are research papers using more sophisticated forms of joint label inference. keras.io/examples/nlp/m…