README.md

Unified Named Entity Recognition as Word-Word Relation Classification

AAAI 2022 paper: Unified Named Entity Recognition as Word-Word Relation Classification

This repository corresponds to the PaddlePaddle implementation of W2NER.

1. Environments

- python (3.8.12)
- cuda (11.4)

2. Dataset

• Conll 2003
• OntoNotes 4.0
• OntoNotes 5.0
• ACE 2004
• ACE 2005
• GENIA
• CADEC
• ShARe13
• ShARe14

3. Preparation

• Download dataset
• Process them to fit the same format as the example in data/
• Put the processed data into the directory data/

4. Training

>> python main.py --config ./config/example.json

5. Evaluation

>> python main.py --ckpt [Pretrained Model Path]

6. Pretrained models

We provide a pretrained model, which can be downloaded from BaiduYun, keyword is: edns. You should unzip and put it in the main directory.