| language: | |
| - en | |
| license: apache-2.0 | |
| size_categories: | |
| - 10K<n<100K | |
| task_categories: | |
| - token-classification | |
| pretty_name: buster | |
| tags: | |
| - finance | |
| configs: | |
| - config_name: default | |
| data_files: | |
| - split: FOLD_1 | |
| path: data/FOLD_1-* | |
| - split: FOLD_2 | |
| path: data/FOLD_2-* | |
| - split: FOLD_3 | |
| path: data/FOLD_3-* | |
| - split: FOLD_4 | |
| path: data/FOLD_4-* | |
| - split: FOLD_5 | |
| path: data/FOLD_5-* | |
| - split: SILVER | |
| path: data/SILVER-* | |
| dataset_info: | |
| features: | |
| - name: document_id | |
| dtype: string | |
| - name: tokens | |
| sequence: string | |
| - name: labels | |
| sequence: string | |
| splits: | |
| - name: FOLD_1 | |
| num_bytes: 9801188 | |
| num_examples: 753 | |
| - name: FOLD_2 | |
| num_bytes: 9719997 | |
| num_examples: 759 | |
| - name: FOLD_3 | |
| num_bytes: 9810761 | |
| num_examples: 758 | |
| - name: FOLD_4 | |
| num_bytes: 9973098 | |
| num_examples: 755 | |
| - name: FOLD_5 | |
| num_bytes: 9831525 | |
| num_examples: 754 | |
| - name: SILVER | |
| num_bytes: 80615802 | |
| num_examples: 6196 | |
| download_size: 20217911 | |
| dataset_size: 129752371 | |
| # Dataset Card for BUSTER | |
| BUSiness Transaction Entity Recognition dataset. | |
| BUSTER is an Entity Recognition (ER) benchmark for entities related to business transactions. It consists of a gold corpus of | |
| 3779 manually annotated documents on financial transactions that were randomly divided into 5 folds, | |
| plus an additional silver corpus of 6196 automatically annotated documents that were created by the model-optimized RoBERTa system. |