HERO: Hierarchical Encoder for Video+Language Omni-representation Pre-training
Open paper
Benchmark
Model
Rank
Results
video-question-answering-on-how2qa
Hero w/ pre-training
#4
Accuracy: 77.75
video-question-answering-on-tvqa
Hero w/ pre-training
#4
Accuracy: 74.24
Rank counts only results with a code link.