Downloads · 30 days
9
2% of all-time downloads
RJ3vans/NPheadTagger
NPheadTagger is a token classification model from RJ3vans. Use it when you need labels on individual words, such as names. It is set up for transformers.
Input a noun phrase (NP) and the model will tag its syntactic head as HEAD. Non-head tokens will be tagged as NA.
Downloads · 30 days
9
2% of all-time downloads
All-time downloads
478
Public
Repo size
2.7 GB
Likes
0
Public
Click a slice to open those files.
.bin1.3 GB · 100%
From the Hugging Face model README
Input a noun phrase (NP) and the model will tag its syntactic head as HEAD. Non-head tokens will be tagged as NA.
The model is intended to tag multiple heads in compound NPs. It is not intended to tag all heads in complex NPs.
The training data is derived from a dataset in which NPs are identified using the Berkeley neural parser which does not identify the heads of syntactic constituents. The training data includes erroneous NPs in which each token is tagged as ERROR.
Examples you might try include:
The bottle of ginger beer
John, Peter, and Mary, who live in Sri Lanka
The man who won the race