C10X/nemotron-hq-1b-0to4k-testv334

Nemotron HQ 1B 0-to-4K Test v334 是一个数据集，其卡片定义了显式的分割映射，包括训练集、评估集和测试集，以便在Hugging Face数据集查看器中分别展示这些分割。数据集文件以parquet格式存储，每个分割包含两个子文件：一个对应短序列（索引0001到2048），另一个对应长序列（索引2049到4096）。Nemotron HQ 1B 0-to-4K Test v334 is a dataset whose dataset card defines explicit split mappings covering training, validation, and test splits, allowing these splits to be displayed separately in the Hugging Face Dataset Viewer. The dataset files are stored in Parquet format, and each split contains two sub-files: one for short sequences with indices ranging from 0001 to 2048, and the other for long sequences with indices ranging from 2049 to 4096.