Datasets:

ArXiv:
License:

Validation split?

#1
by cfierro - opened

In the original XLM-R paper and in other retrains of XLM-R (https://arxiv.org/abs/2205.11758) they mention a validation split of cc100, do we know how the validation split was selected? Could we include it in this dataset?

Statistical and Neural Machine Translation org

Hi @cfierro , we are sorry but we can't find any mention to a validation split of cc100 in the original paper.

albertvillanova changed discussion status to closed

Sign up or log in to comment