corniclr25/stack-mined-javascript-v1

该数据集包含四个主要特征：查询（query）、文档（document）、负面样本（negatives）和元数据（metadata）。元数据进一步包含目标（objective）结构，该结构又包含配对（paired）、自我（self）和三元组（triplet）三个子结构。数据集分为一个训练集（train），包含2,334,784个样本，总大小为29,123,533,349字节。下载大小为11,806,323,547字节。

The dataset contains four main features: query, document, negatives, and metadata. The metadata further includes an objective structure, which contains three substructures: paired, self, and triplet. The dataset is divided into one training set (train) containing 2,334,784 samples, with a total size of 29,123,533,349 bytes. The download size is 11,806,323,547 bytes.