Skip to content

Latest commit

 

History

History
9 lines (6 loc) · 440 Bytes

seq_bio.md

File metadata and controls

9 lines (6 loc) · 440 Bytes

seq_bio task-type

back to main README

The usage of this task type is similar as seq. However, this task type assumes that your data is in BIO format, and also guarantees that the output will be in valid BIO format. Under the hood it is using a masked CRF, which disallows ill-formed output. Performance can be expected to be higher for span labeling tasks (like NER) compared to the standard seq decoder.