lightonai/embeddings-pre-training

这是一个大型的文本嵌入模型预训练数据集，包含多样化的对比数据，旨在开发最先进的文本嵌入模型。该数据集主要以英语为主，同时包含几个法语数据集，以支持双语文本和跨语言研究。

This is a large-scale dataset of diverse, contrastive pre-training data for developing state-of-the-art text embedding models. The collection is primarily in English, with the inclusion of several French datasets to facilitate bilingual and cross-lingual research.