| dataset_info: | |
| features: | |
| - name: sentence | |
| dtype: string | |
| - name: target | |
| dtype: | |
| class_label: | |
| names: | |
| '0': eight | |
| '1': five | |
| '2': four | |
| '3': nine | |
| '4': 'no' | |
| '5': one | |
| '6': seven | |
| '7': six | |
| '8': ten | |
| '9': three | |
| '10': two | |
| '11': zero | |
| splits: | |
| - name: train | |
| num_bytes: 659600 | |
| num_examples: 8320 | |
| - name: validation | |
| num_bytes: 81438 | |
| num_examples: 1029 | |
| - name: test | |
| num_bytes: 86986 | |
| num_examples: 1095 | |
| download_size: 438887 | |
| dataset_size: 828024 | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: train | |
| path: data/train-* | |
| - split: validation | |
| path: data/validation-* | |
| - split: test | |
| path: data/test-* | |
| source_datasets: | |
| - INK-USC/numer_sense | |
| # numer_sense | |
| NumerSense masked number words (0-10), with deterministic 80/10/10 splits. | |
| Original data: [INK-USC/numer_sense](https://huggingface.co/datasets/INK-USC/numer_sense). Repackaged as parquet for [tasksource](https://github.com/sileod/tasksource) by [scripts/upload_repackaged.py](https://github.com/sileod/tasksource/blob/main/scripts/upload_repackaged.py). | |