🐛 Trying to use ROUGE metric : pyarrow.lib.ArrowInvalid: Column 1 named references expected length 534 but got length 323 about datasets HOT 5 CLOSED

This is because add takes as input a batch of elements and you provided only one. I think we should have add for one prediction/reference and add_batch for a batch of predictions/references. This would make it more coherent with the way we use Arrow.

Let me do this change

from datasets.

Can you share your data files or a minimally reproducible example?

from datasets.

Sure, here is a Colab notebook reproducing the error.

ArrowInvalid: Column 1 named references expected length 36 but got length 56

from datasets.

Thanks for noticing though. I was mainly used to do .compute directly ^^

from datasets.

Thanks @lhoestq it works :)

from datasets.