使用哈工大，BERT-wwm，中文bert，在20w中文人物关系数据上的准确率达到0.97
=== Epoch 0 train === 100%|██████████████████████████████████████████████████████████████████| 3094/3094 [40:12<00:00, 1.28it/s, acc=0.773, loss=0.687] === Epoch 0 val === 100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:06<00:00, 2.42it/s, acc=0.934] Best ckpt and saved. === Epoch 1 train === 100%|██████████████████████████████████████████████████████████████████| 3094/3094 [38:17<00:00, 1.35it/s, acc=0.923, loss=0.235] === Epoch 1 val === 100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:05<00:00, 2.78it/s, acc=0.972] Best ckpt and saved. === Epoch 2 train === 100%|██████████████████████████████████████████████████████████████████| 3094/3094 [22:43<00:00, 2.27it/s, acc=0.961, loss=0.121] === Epoch 2 val === 100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:05<00:00, 2.71it/s, acc=0.986] Best ckpt and saved. Best acc on val set: 0.986000 100%|██████████████████████████████████████████████████████████████████████████████████| 16/16 [00:06<00:00, 2.64it/s, acc=0.986] Accuracy on test set: 0.986
model.infer({'text': '场照片事后将发给媒体，避免采访时出现混乱，[3]举行婚礼侯佩岑黄伯俊婚纱照2011年4月17日下午2点，70名亲友见 证下，侯佩', 'h': {'pos': (28, 30)}, 't': {'pos': (31, 33)}})
('夫妻', 0.9995878338813782)
model.infer({'text': '及他们的女儿小苹果与汪峰感情纠葛2004年，葛荟婕在欧洲杯期间录制节目时与汪峰相识并相恋，汪峰那首《我如此爱你', 'h': {'pos': (10, 11)}, 't': {'pos': (22, 24)}})
('情侣', 0.9992896318435669)
model.infer({'text': '14日，彭加木的侄女彭丹凝打通了彭加木儿子彭海的电话，“堂哥已经知道了，他说这些年传得太多，他不相信是真的', 'h': {'pos': (4, 6)}, 't': {'pos': (22, 21)}})
('父母', 0.8954808712005615)
model.infer({'text': '名旦吴菱仙是位列“同治十三绝”的名旦时小福的弟子，算得梅兰芳的开蒙老师，早年曾搭过梅巧玲的四喜班，旧谊', 'h': {'pos': (2, 4)}, 't': {'pos': (27, 29)}})
('师生', 0.996309220790863)
1.bert模型下载：在./pretrain/下面放置chinese_wwm_pytorch模型，下载地址：https://github.com/ymcui/Chinese-BERT-wwm
2.数据下载：在./benchmark/people-relation/下执行gen.py,生产中文人物关系数据，具体脚本中有说明。
3.配置环境变量：vim ~/.bash_profile 添加
# openNRE
export openNRE=项目位置
如果自己训练了tensorflow 的bert，可以通过https://github.com/huggingface/transformers 里面的convert_bert_original_tf_checkpoint_to_pytorch.py 脚本转换为pytorch版。
踩坑：
1.安装tensorflow 2.0，最终用的都是PyTorch模型，但TensorFlow也得安装
2.构造checkpoint文件
3.报错：Embedding' object has no attribute 'shape' ，解决：将报错位置assert那几行直接删除
以下是原工程内容
OpenNRE is an open-source and extensible toolkit that provides a unified framework to implement relation extraction models. This package is designed for the following groups:
- New to relation extraction: We have hand-by-hand tutorials and detailed documents that can not only enable you to use relation extraction tools, but also help you better understand the research progress in this field.
- Developers: Our easy-to-use interface and high-performance implementation can acclerate your deployment in the real-world applications. Besides, we provide several pretrained models which can be put into production without any training.
- Researchers: With our modular design, various task settings and metric tools, you can easily carry out experiments on your own models with only minor modification. We have also provided several most-used benchmarks for different settings of relation extraction.
- Anyone who need to submit an NLP homework to impress their professors: With state-of-the-art models, our package can definitely help you stand out among your classmates!
This package is mainly contributed by Tianyu Gao, Xu Han, Shulian Cao, Lumin Tang, Yankai Lin, Zhiyuan Liu
Relation extraction is a natural language processing (NLP) task aiming at extracting relations (e.g., founder of) between entities (e.g., Bill Gates and Microsoft). For example, from the sentence Bill Gates founded Microsoft, we can extract the relation triple (Bill Gates, founder of, Microsoft).
Relation extraction is a crucial technique in automatic knowledge graph construction. By using relation extraction, we can accumulatively extract new relation facts and expand the knowledge graph, which, as a way for machines to understand the human world, has many downstream applications like question answering, recommender system and search engine.
A good research work is always accompanied by a thorough and faithful reference. If you use or extend our work, please cite the following paper:
@inproceedings{han2019opennre,
title={OpenNRE: An Open and Extensible Toolkit for Neural Relation Extraction},
author={Han, Xu and Gao, Tianyu and Yao, Yuan and Ye, Deming and Liu, Zhiyuan and Sun, Maosong },
booktitle={Proceedings of EMNLP},
year={2019}
}
It's our honor to help you better explore relation extraction with our OpenNRE toolkit!
If you want to learn more about neural relation extraction, visit another project of ours (NREPapers).
You can refer to our document for more details about this project.
We are now working on deploy OpenNRE as a Python package. Coming soon!
Clone the repository from our github page (don't forget to star us!)
git clone https://github.com/thunlp/OpenNRE.gitIf it is too slow, you can try
git clone https://github.com/thunlp/OpenNRE.git --depth 1
Then install all the requirements:
pip install -r requirements.txt
Then install the package with
python setup.py install
If you also want to modify the code, run this:
python setup.py develop
Note that we have excluded all data and pretrain files for fast deployment. You can manually download them by running scripts in the benchmark and pretrain folders. For example, if you want to download FewRel dataset, you can run
bash benchmark/download_fewrel.shAdd OpenNRE directory to the PYTHONPATH environment variable, or open a python session under the OpenNRE folder. Then import our package and load pre-trained models.
>>> import opennre
>>> model = opennre.get_model('wiki80_cnn_softmax')Note that it may take a few minutes to download checkpoint and data for the first time. Then use infer to do sentence-level relation extraction
>>> model.infer({'text': 'He was the son of Máel Dúin mac Máele Fithrich, and grandson of the high king Áed Uaridnach (died 612).', 'h': {'pos': (18, 46)}, 't': {'pos': (78, 91)}})
('father', 0.5108704566955566)You will get the relation result and its confidence score.
For higher-level usage, you can refer to our document.
If you want to receive our update news or take part in discussions, please join our Google Group