Twitter4sse

🚀 Twitter4SSE

Twitter4SSE 模型可将文本映射到 768 维的密集嵌入向量，这些向量能够对语义相似度进行编码。该模型在 Twitter 数据集上使用多负例排序损失（MNRL）进行训练。它基于 BERTweet 进行初始化，并使用 Sentence-transformers 库进行训练。

🚀 快速开始

本模型借助 sentence-transformers 库使用起来更加便捷。

📦 安装指南

pip install -U sentence-transformers

💻 使用示例

基础用法

from sentence_transformers import SentenceTransformer
sentences = ["This is the first tweet", "This is the second tweet"]

model = SentenceTransformer('digio/Twitter4SSE')
embeddings = model.encode(sentences)
print(embeddings)

若不使用 sentence-transformers 库，请参考 此仓库 获取在 Huggingface 上使用 Sentence Transformers 的详细说明。

📚 详细文档

引用与作者

官方论文 Exploiting Twitter as Source of Large Corpora of Weakly Similar Pairs for Semantic Sentence Embeddings 将在 2021 年的 EMNLP 会议上发表。更多详细信息即将公布。

@inproceedings{di-giovanni-brambilla-2021-exploiting,
    title = "Exploiting {T}witter as Source of Large Corpora of Weakly Similar Pairs for Semantic Sentence Embeddings",
    author = "Di Giovanni, Marco  and
      Brambilla, Marco",
    booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2021",
    address = "Online and Punta Cana, Dominican Republic",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.emnlp-main.780",
    pages = "9902--9910",
}

官方代码可在 GitHub 上获取。

📄 许可证

本项目采用 Apache-2.0 许可证。