Self-GRIT/PILE_Wikipedia_Pretraining_subset_100k
收藏官方服务:
资源简介:
--- dataset_info: features: - name: text dtype: string splits: - name: 50k_part0 num_bytes: 103754524 num_examples: 50000 - name: 50k_part1 num_bytes: 104483416 num_examples: 50000 - name: valid num_bytes: 1072820 num_examples: 500 download_size: 305556399 dataset_size: 209310760 configs: - config_name: default data_files: - split: 50k_part0 path: data/50k_part0-* - split: 50k_part1 path: data/50k_part1-* - split: valid path: data/valid-* ---
提供机构:
Self-GRIT


