Is the evaluation data available / can you make it available? Sorry if it already is and I missed it.
Specifically referring to this part in the paper:
We construct test sets from the WMT News Translation test sets (Barrault et al., 2020), which provides for decent-size test sets in many languages. We manually corrected all sentence segmentations. While some sets were already well segmented, some more recent years were extremely under-segmented.
Having the evaluation data as used in your paper would make it much easier to build on your work. Thanks!
Is the evaluation data available / can you make it available? Sorry if it already is and I missed it.
Specifically referring to this part in the paper:
Having the evaluation data as used in your paper would make it much easier to build on your work. Thanks!