Bertoverflow Stackoverflow Github

🚀 BERTOverflow

BERTOverflow is a pre - trained model based on the BERT - base architecture, trained on 152 million sentences from StackOverflow's 10 - year archive, which can be used for code and named entity recognition.

🚀 Quick Start

✨ Features

We pre - trained the BERT - base model on 152 million sentences from the StackOverflow's 10 - year archive. More details of this model can be found in our ACL 2020 paper: Code and Named Entity Recognition in StackOverflow.

💻 Usage Examples

Basic Usage

from transformers import *
import torch

tokenizer = AutoTokenizer.from_pretrained("lanwuwei/BERTOverflow_stackoverflow_github")
model = AutoModelForTokenClassification.from_pretrained("lanwuwei/BERTOverflow_stackoverflow_github")

📄 License

BibTeX entry and citation info

@inproceedings{tabassum2020code,
    title={Code and Named Entity Recognition in StackOverflow},
    author={Tabassum, Jeniya  and Maddela, Mounica  and Xu, Wei and Ritter, Alan },
    booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL)},
    url={https://www.aclweb.org/anthology/2020.acl-main.443/}
    year = {2020},
}