clulab / releases

31 stars 14 forks source link

Wassa2023 data and training/eval notebooks #21

Closed maxaalexeeva closed 1 year ago

maxaalexeeva commented 1 year ago

@MihaiSurdeanu, there are some columns in some tsvs that are not strictly necessary, e.g., tokenized sentence. Also, there might be a column or two that have the same information but are named slightly different (eg tokenized_sentence vs tok_sents)---is it worth it going through them to check for consistency? Arguments against renaming/deleting columns are a. it takes time (which is fine) and b. sometimes editing files locally messes up tsvs.

MihaiSurdeanu commented 1 year ago

I think we should keep the data as close as possible to the paper. So, I recommend to keep it as is. Merge?

maxaalexeeva commented 1 year ago

Yes, I think providing too much info is better than not enough. Should be ok to merge then. Thanks!

MihaiSurdeanu commented 1 year ago

Merged! Thanks!