https://github.com/shashwath94/relation-extraction-using-cnn
A Chainer implementation of a Convolutional Network model for relation classification in the SemEval Task 8 dataset. This model performs Multi-Way Classification of Semantic Relations Between Pairs of Nominals in the SemEval 2010 task 8 dataset.
https://github.com/shashwath94/relation-extraction-using-cnn
chainer cnn-text-classification convolutional-neural-networks deep-learning python3
Last synced: 11 months ago
JSON representation
A Chainer implementation of a Convolutional Network model for relation classification in the SemEval Task 8 dataset. This model performs Multi-Way Classification of Semantic Relations Between Pairs of Nominals in the SemEval 2010 task 8 dataset.
- Host: GitHub
- URL: https://github.com/shashwath94/relation-extraction-using-cnn
- Owner: shashwath94
- Created: 2017-11-15T14:58:29.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2018-01-16T20:04:21.000Z (over 8 years ago)
- Last Synced: 2024-12-16T03:04:49.092Z (over 1 year ago)
- Topics: chainer, cnn-text-classification, convolutional-neural-networks, deep-learning, python3
- Language: Python
- Homepage:
- Size: 497 KB
- Stars: 17
- Watchers: 3
- Forks: 7
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Relation-Extraction-using-CNN
A Chainer implementation of a Convolutional Network model for relation classification in the SemEval Task 8 dataset. This model performs Multi-Way Classification of Semantic Relations Between Pairs of Nominals in the SemEval 2010 task 8 dataset.
The CNN model is inspired by Relation Classification via Convolutional Deep Neural Network
Requirements
Requirements
1. Python3
2. Chainer
3. vsmlib
4. numpy
5. Word Embeddings (It can be downloaded from https://nlp.stanford.edu/projects/glove/, the Stanford NLP group has a bunch of open source pre-trained Glove embeddings or you can use your own embeddings. Just specify the path in config.yaml)
Dataset
The Semeval 2010 task 8 dataset can be downloaded from https://docs.google.com/document/d/1QO_CnmvNRnYwNWu1-QCAeR5ToQYkXUqFeAJbdEhsq7w/preview. A small subset of the data has been provided to get you started. The format of the data is as follows:
```
Component-Whole(e2,e1) 12 15 The system as described above has its greatest application in an arrayed configuration of antenna elements .
```
The first part is the label ie, the relation between the nominals present at index 12 and 15 respectively.
Configuration parameters
All the config parameters and the hyperparameters of the model can be specified in the config.yaml file.
Train the model
```
python3 main.py config.yaml