This PR significantly cleans up the BC5CDR script. It also adds a couple of QOL improvements, like better CLI documentation. Also, previously one would have to type seq2rel-ds preprocess bc5cdr bc5cdr, but this has been fixed (it's now seq2rel-ds preprocess bc5cdr)
TODO
[x] Add tests for BC5CDR
[x] Figure out why the number of processed examples != the expected value Tracking this in #13.
Overview
This PR significantly cleans up the BC5CDR script. It also adds a couple of QOL improvements, like better CLI documentation. Also, previously one would have to type
seq2rel-ds preprocess bc5cdr bc5cdr, but this has been fixed (it's nowseq2rel-ds preprocess bc5cdr)TODO
[x] Figure out why the number of processed examples != the expected valueTracking this in #13.