| dataset_info: | |
| features: | |
| - name: text | |
| dtype: string | |
| - name: meta | |
| struct: | |
| - name: pile_set_name | |
| dtype: string | |
| splits: | |
| - name: train | |
| num_bytes: 542456309 | |
| num_examples: 100000 | |
| download_size: 202975613 | |
| dataset_size: 542456309 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| ## Dataset Creation Process | |
| These subsets were created by streaming over the rows from [`monology/pile-uncopyrighted`](https://huggingface.co/datasets/monology/pile-uncopyrighted) and filtering by the meta column. Each subset is generally limited to the first 100,000 qualifying rows encountered. | |