This repository provides a tool box for the Dynamic Named Entity Recognition framework.
Data can be found at this link, or manually search
for :
https://drive.google.com/file/d/1u8zQZV1gHyzt3bPF4RGQSt_dciRvVKLA/view?usp=sharing
Data should be placed in the data\ directory to use pre defined scripts.
Three environment variable soudle be set :
$DATA_ROOT: Directory holding the data (the directoryDNER_datasetsshould be placed under it)$LOG_ROOT: Directory holding the logged data (mainlytensorboardfiles)$INFERENCE_ROOT: Directory holding the model output data
The repo is organized as follow :
- run.py : main python script for training models
- script\ : contains sh script with predefined configuration for model training
- source\ : source files
- data\ : directory holding training data
- requirement.txt : requirements file for module importation