Datasets:

ncbi_disease

like 37

There are a couple of empty examples in the dataset. For future reference, you can drop these empty rows by:
data = data.filter(lambda example: len(example["ner_tags"]) > 0)