Sacrebleu scoring method always returns score of 0?

Course Q&A Natural Language Processing NLP with Attention Models
1

Hi, in the notebook for BLEU scoring, there’s a separate call to the external library sacrebleu to use as a comparison with the custom BLEU scoring method, but the actual results of the sacrebleu scoring calls are always 0.0

Can somebody verify this isn’t expected behavior?

1 Like
2

This topic is closed.