may ohta

> It seem the register is closed Do you mean the registration of CoNLL 2016 Shared Task? This project assumes that you already have the CoNLL 2016 data. At this...

Hello @vmenan, I think [`collate_fn()`](https://github.com/joeynmt/joeynmt/blob/cdc4d03d430a1b0f29793a0d95743c5e72ae2f6c/joeynmt/datasets.py#L175) is an appropriate place to extend. What type of input format do you use? plain text, tsv, or huggingface's dataset? If you could provide some...

Hi @vmenan, Ah, ok, now I understand your project better. So, you'd like to compute the loss separately for each dataset, right? Then you need to change the training manager...

Hello, Which processed data exactly, do you mean? We are responsible only for the data linked from https://github.com/joeynmt/joeynmt repo, we have no idea (and no control) on anything outside of...