gucorpling / amalgum

English web corpus with 4M tokens and several annotation types
25 stars 6 forks source link

Add seq2set entity recognizer #19

Closed yilunzhu closed 2 years ago

yilunzhu commented 2 years ago

The module's input and output are both conllu formats. For the entity recognizer, only entity type and entity ids are added in the output. The module is tested on local environment with CPU and GCloud with GPU. Also, dependencies are added to requirements.txt