stephantulkens/swim-ir-query-gte-modernbert-pooled

这是一个使用阿里巴巴-NLP/gte-modernbert-base模型在nthakur/swim-ir-monolingual数据集上生成的嵌入数据集。每个示例都包含原始文本和其相应的768维嵌入。该数据集适用于大规模蒸馏、检索和相似性搜索等任务。README还提供了关于数据集分割的信息，特别是包含501,538个示例的train分割。此外，它还提供了关于如何处理嵌入的建议，例如将嵌入截断以适应模型的token长度限制。

This is a dataset of embeddings created from the Alibaba-NLP/gte-modernbert-base model on the nthakur/swim-ir-monolingual dataset. Each example in the dataset includes the original text and its corresponding 768-dimensional embedding. The dataset is intended for tasks such as large-scale distillation, retrieval, and similarity search. The README also provides information about the datasets splits, specifically the train split which contains 501,538 examples. Additionally, it offers recommendations for working with the embeddings, such as truncating the embeddings to fit within the models token length limit.