A quick-and-easy way to make ULMFiT work for multi-label problems

Thanks.
I’m using the output from an RNN encoder as features, in addition to continuous and categorical features, of a structured classifier. I’ll post the code here when it’s ready in case it can help anyone.

2 Likes

How training data is supposed to be???
is it shoukd be like this for 3 labels??

label1 label2 label3 text