DUDE competition train - validation - test splits ground truth

This JSON file contains the ground truth annotations for the train and validation set of the DUDE competition (https://rrc.cvc.uab.es/?ch=23&amp;com=tasks) of ICDAR 2023 (https://icdar2023.org/). <strong>V1.0.7 release</strong>: 41454 annotations for 4974 documents (train-validation-test) <pre>DatasetDict({ train: Dataset({ features: ['docId', 'questionId', 'question', 'answers', 'answers_page_bounding_boxes', 'answers_variants', 'answer_type', 'data_split', 'document', 'OCR'], num_rows: 23728 }) val: Dataset({ features: ['docId', 'questionId', 'question', 'answers', 'answers_page_bounding_boxes', 'answers_variants', 'answer_type', 'data_split', 'document', 'OCR'], num_rows: 6315 }) test: Dataset({ features: ['docId', 'questionId', 'question', 'answers', 'answers_page_bounding_boxes', 'answers_variants', 'answer_type', 'data_split', 'document', 'OCR'], num_rows: 11411 }) }) ++update on answer_type +++formatting change to answers_variants ++++stricter check on answer_variants &amp; rename annotations file <strong>+ blind test set (no ground truth answers provided)</strong> </pre>