Understanding NER F1 score seqeval calculation

[KrishnanJothi]

Hello everyone,

Please see the below image, which consists of the model prediction and target sentences with labels. The number on the right is the f1 micro score. The following code is used to generate the metric:

f1_score(valid_tags,pred_tags, average='micro', mode='strict', scheme=IOB2)

Can anyone please explain how it is calculated?