/CCKS-2021-event-detection

基于BERT的事件检测模型，适用于MAVEN、ACE等数据集。

Primary LanguagePython

CCKS-2021-event-detection

We only offer baseline model on CCKS 2021：通用细粒度事件检测. F1-score is at least 0.68.

本项目实现基于BERT的事件检测模型，后续会更新BERT+CNN、BERT+BiLSTM等模型。

This code is the implementation for DMBERT model. The implementations are based on Huggingface's Transformers, especially its example for the multiple-choice task.

Requirements

  • python==3.6.9

  • torch==1.2.0

  • transformers==2.8.0

  • sklearn==0.20.2

Usage

On MAVEN:

  1. Download MAVEN data files.
  2. Run main.slurm for training and evaluation on the devlopment set, and geting predictions on the test set (dumped to results.jsonl).

See the two scripts for more details.

On ACE

  1. Preprocess ACE 2005 dataset as in this repo.
  2. Run main.slurm for training and evaluation.