| dataset_info: | |
| features: | |
| - name: content | |
| dtype: string | |
| - name: markdown | |
| dtype: string | |
| - name: url | |
| dtype: string | |
| splits: | |
| - name: train | |
| num_bytes: 82642012 | |
| num_examples: 2067 | |
| download_size: 34862377 | |
| dataset_size: 82642012 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| It is what it says on the tin, 2k websites that don't run javascript (aka: probably really old, and simple!). Filtering needs to be done to take out the defunct sites and error messages. I don't expect this is a hard task. |