Hey what are some ways I can use ulmfit to get an embedding like representation for a sentence. Would taking the context vector obtained after the last LSTM later help as a good representation of a sentence?. I want to use it to check for similar sentences
1 Like