/Named-Entity-Recognition-NER-using-spaCy-ver3

Primary LanguageJupyter NotebookMIT LicenseMIT

Named Entity Recognition NER using spaCy ver3

Named Entity Recognition using CLI training process

Repo includes folowing config files to train models:

-config_demo_acc (lg model based)
-config_demo_efficiency (train from empty)
-config_demo_transformer (trf model based)
  • config files use wandb as logger
  • details about training and tests processes provided in NER_de_training.ipynb notebook

Model usage

-trained models include best checkpoint (checkpoint setting can be changed in config files)
-to run model use "spacy.load(path_to_model)"
-in the repo provided smallest and simpliest model

Report plots

drawing

drawing

Prerequisites

-spaCy
-sklearn

License

This project is licensed under the MIT License - see the LICENSE file for details