CR-CNN-relation-extraction

Implementation of Classifying Relations by Ranking with Convolutional Neural Networks.

Environment Requirements

• python 3.6
• pytorch 1.3.0

Data

• SemEval2010 Task8 [paper]
• glove.6B.300d.txt

Usage

1. Download the embedding and decompress it into the embedding folder.
2. Run the following the commands to start the program.

python run.py

More details can be seen by python run.py -h.

3. You can use the official scorer to check the final predicted result.

perl semeval2010_task8_scorer-v1.2.pl proposed_answer.txt predicted_result.txt >> result.txt

Result

The result of my version and that in paper are present as follows:

paper	my version
0.841	0.8327

The training log can be seen in train.log and the official evaluation results is available in result.txt.

Note:

• Some settings are different from those mentioned in the paper.
• No validation set used during training.
• Although I try to set random seeds, it seems that the results of each run are a little different.

Reference Link

• https://github.com/lawlietAi/CRCNN-Pytorch
• https://github.com/AnblueWang/cnn-relation-extraction-with-ranking-loss