ssuresh/nemo-stage1-50M-samples
收藏Hugging Face2025-10-28 更新2025-11-15 收录
下载链接:
https://hf-mirror.com/datasets/ssuresh/nemo-stage1-50M-samples
下载链接
链接失效反馈官方服务:
资源简介:
NeMo Stage1预训练数据集包含约5000万条英文文本样本,用于NeMo模型的预训练(阶段1)。数据集以分块方式组织,以便高效加载和处理。每个样本包含一个唯一的ID和文本内容,以JSONL格式存储。总共有47个块,每个块大约包含100万个样本。
The NeMo Stage1 Pretraining Dataset contains approximately 50 million English text samples for NeMo model pretraining (Stage 1). The dataset is organized in chunks for efficient loading and processing. Each sample includes a unique ID and text content, stored in JSONL format. There are a total of 47 chunks, with each chunk containing about 1,000,000 samples.
提供机构:
ssuresh



