OpenNMT-tf
OpenNMT-tf copied to clipboard
[Multi-Task] Word Feature Output in Target Side
OpenNMT Lua version supports extra word features concatenated with token with "|" delimiter and the decoder can also make predictions of word features. I'm wondering if Tensorflow version has this functionality?
Ok, I found that we can incorporate word features with parallel inputs. But it seems like we cannot make the decoder predict those features like Lua version? (a.k.a multi-task learning)
You are correct, additional word features are only supported on the source side using parallel inputs.