codelion/qwen-fineweb-distilled

该数据集包含了多个特征字段，其中包括教师嵌入维度（teacher_embedding_dim）、是否为元数据（is_metadata）、输入ID序列（input_ids）、序列长度（sequence_lengths）、精确值序列（exact_values）、系数序列（coeffs）和打包索引（packed_indices）。数据集分为训练集（train），包含55个示例，总大小为8391068355字节。数据集的下载大小为1971033491字节。数据集提供了默认配置，其中指定了训练数据的文件路径。

The dataset includes multiple feature fields such as teacher embedding dimension (teacher_embedding_dim), whether it is metadata (is_metadata), input ID sequence (input_ids), sequence length (sequence_lengths), exact value sequence (exact_values), coefficient sequence (coeffs), and packed indices (packed_indices). The dataset is split into a training set (train) containing 55 examples, with a total size of 8391068355 bytes. The download size of the dataset is 1971033491 bytes. The dataset provides a default configuration, which specifies the file path for the training data.