search

Franck-Dernoncourt / NeuroNER

Named-entity recognition using neural networks. Easy-to-use and state-of-the-art results.
http://neuroner.com
MIT License
1.69k stars 476 forks source link

Character embedding question #135

Closed ntzky closed 5 years ago

ntzky commented 5 years ago

How did you represent a character as a vector? I found the file "tensorboard_metadata_characters.tsv", did you use 1-hot encoding?

Franck-Dernoncourt commented 5 years ago

We used character embeddings (default dimension: 25). You can change the dimension in parameters.ini by changing the value ofcharacter_embedding_dimension.

See https://doi.org/10.1093/jamia/ocw156 for the model description.