According to the announcement on the website, the hidden test dataset will be used to evaluate the submission. And as the data shown in the provided public_test_mmsys_final.json, including "maskrcnn_bboxes", "caption1/2_modified" and "bert_base/large_score", will those also be provided in the hidden test dataset?
Thanks for answering the questions.
Sincerely,
Ting-Hsuan