StrucTexTv2: Masked Visual-Textual Prediction for Document Image Pre-training

Benchmark Model Rank Results
semantic-entity-labeling-on-funsdStrucTexTv2 (large)#5F1: 91.82
semantic-entity-labeling-on-funsdStrucTexTv2 (small)#8F1: 89.23
document-image-classification-on-rvl-cdipStrucTexTv2 (large)#14Accuracy: 94.62%Parameters: 238M
document-image-classification-on-rvl-cdipStrucTexTv2 (small)#17Accuracy: 93.4%Parameters: 28M