fineweb10b-gpt2-subset
收藏官方服务：
资源简介：
1. All token values range from 0-50256 (GPT-2 vocabulary size) 2. Files use little-endian byte order 3. Recommend reading training files sequentially for best performance 4. Validation file is for model evaluation and should not be used for training
提供机构：
maas创建时间：
2025-12-23
